Safety protection method and electronic device for electric two-wheeler

Abstract

Embodiments of the present application provide a kind of electric two-wheeled vehicle safety protection method and electronic equipment, wherein the method comprises: in response to the power-on of electric two-wheeled vehicle, the vehicle sensor parameters of electric two-wheeled vehicle are collected, and the environmental data of the environment where electric two-wheeled vehicle is located is received and issued by cloud server;Based on vehicle sensor parameters and environmental data, collision risk prediction is carried out on electric two-wheeled vehicle;In response to predicting that electric two-wheeled vehicle exists collision risk, at least one safety protection action is executed, and trigger signal is sent to wearable safety terminal;At least one safety protection action includes at least one of sound and light early warning, instrument display and vehicle-mounted air bag point explosion;Trigger signal is used to indicate that at least one safety response action is executed by wearable safety terminal, and at least one safety response action includes at least one of executing somatic early warning, wearable air bag point explosion and sound and light early warning. Through the present application, the problem that wearable safety terminal cannot early warning is solved.

Description

Technical Field

[0001] This application relates to the field of two-wheeled vehicle safety protection, and more specifically, to a safety protection method and electronic device for an electric two-wheeled vehicle. Background Technology

[0002] Electric two-wheeled vehicles (such as two-wheeled motorcycles) are the mainstream means of transportation for daily commuting and leisure travel in China. However, their open riding structure leads to a much higher rider injury and death rate in accidents compared to four-wheeled vehicles. Currently, riders mainly rely on wearing independent wearable safety devices for safety protection.

[0003] However, existing wearable safety terminals can only be triggered after a collision / crash occurs. They cannot obtain risk signals from the vehicle's sensing unit in advance, cannot issue warnings to riders before a collision occurs, and cannot reduce the probability of accidents. Summary of the Invention

[0004] This application provides a safety protection method and electronic device for electric two-wheeled vehicles, which at least solves the technical problem that wearable safety terminals cannot provide early warnings in related technologies.

[0005] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided, comprising: in response to the electric two-wheeled vehicle being powered on, collecting vehicle sensor parameters of the electric two-wheeled vehicle and receiving environmental data of the environment in which the electric two-wheeled vehicle is located from a cloud server; predicting the collision risk of the electric two-wheeled vehicle based on the vehicle sensor parameters and the environmental data; in response to the prediction that the electric two-wheeled vehicle has a collision risk, executing at least one safety protection action and sending a trigger signal to a wearable safety terminal; the at least one safety protection action includes at least one of: audible and visual warning, instrument display, and vehicle airbag deployment; the trigger signal is used to instruct the wearable safety terminal to execute at least one safety response action, the at least one safety response action including: executing a body-sensing warning, wearable airbag deployment, and audible and visual warning.

[0006] According to another aspect of the embodiments of this application, a computer-readable storage medium is also provided, wherein a computer program is stored therein, wherein the computer program is configured to perform the steps in any of the above method embodiments when executed by a processor.

[0007] According to another aspect of the embodiments of this application, a computer program product or computer program is provided, the computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, causing the computer device to perform the steps in any of the method embodiments described above.

[0008] According to another aspect of the embodiments of this application, an electronic device is also provided, including a memory and a processor, wherein the memory stores a computer program, and the processor is configured to perform the steps of any of the above method embodiments through the computer program.

[0009] Through this application, electric two-wheelers can predict collision risks by collecting vehicle sensor parameters and environmental data. When a collision risk is predicted, safety protection actions are executed and a trigger signal is sent to a wearable safety terminal. This realizes pre-event graded early warning and collaborative protection, solving the problems of existing wearable safety terminals that lack pre-event risk warning capabilities, can only trigger after the event, and cannot avoid accidents in advance. This improves the accuracy of risk identification and the timeliness of protection. Attached Figure Description

[0010] Figure 1 This is a schematic diagram illustrating an application scenario of a safety protection method for an electric two-wheeled vehicle according to an embodiment of this application;

[0011] Figure 2 This is a structural diagram of an optional electric two-wheeler according to an embodiment of this application;

[0012] Figure 3 This is a structural diagram of an optional cloud server according to an embodiment of this application;

[0013] Figure 4 This is a structural diagram of another optional wearable security terminal according to an embodiment of this application;

[0014] Figure 5 This is a structural diagram of another optional mobile terminal according to an embodiment of this application;

[0015] Figure 6 This is a schematic diagram of another optional safety protection method for electric two-wheeled vehicles according to an embodiment of this application;

[0016] Figure 7 This is a schematic diagram of another optional safety protection method for electric two-wheeled vehicles according to an embodiment of this application;

[0017] Figure 8 This is an optional bidirectional redundant branch according to an embodiment of this application;

[0018] Figure 9 This is a schematic diagram of another optional safety protection method for electric two-wheeled vehicles according to an embodiment of this application;

[0019] Figure 10 This is a schematic diagram of another optional safety protection method for electric two-wheeled vehicles according to an embodiment of this application;

[0020] Figure 11 This is a timing diagram of bidirectional encrypted communication between an optional electric two-wheeler and a wearable security terminal according to an embodiment of this application;

[0021] Figure 12 This is a timing diagram of bidirectional encrypted communication between an alternative embodiment of this application and a wearable security terminal;

[0022] Figure 13 This is a schematic diagram of another optional safety protection method for electric two-wheeled vehicles according to an embodiment of this application;

[0023] Figure 14 This is a flowchart illustrating the forced activation of the security protection function of an optional wearable security terminal according to an embodiment of this application;

[0024] Figure 15 This is a flowchart illustrating the forced shutdown of the security protection function of an optional wearable security terminal according to an embodiment of this application. Detailed Implementation

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0026] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided. Optionally, in this embodiment, the above-described safety protection method for an electric two-wheeled vehicle may be applied, but is not limited to, to applications such as... Figure 1 The hardware environment shown includes an electric two-wheeler 102, a cloud server 104, a wearable security terminal 106, and a preset channel 108. The electric two-wheeler 102 communicates with the wearable security terminal 106 via an encrypted two-way communication link; the electric two-wheeler 102 connects to the cloud server 104 via wireless communication technology; the cloud server 104 connects to the preset channel 108 via wireless communication technology; and the preset channel 108 also connects to the electric two-wheeler 102 via wireless communication technology. Here, wireless communication technology refers to the type of wireless communication technology, including but not limited to Wi-Fi, P2P, TCP / IP, MQTT, Bluetooth, LoRa communication, and other wireless connection technologies.

[0028] Electric two-wheeled vehicles 102 refer to motorized or non-motorized two-wheeled vehicles that use electricity as the main driving source, have two or more wheels, and are driven by an electric motor. For example, electric two-wheeled vehicles 102 include electric skateboards, electric bicycles, electric motorcycles, electric mopeds, three-wheeled motorcycles, all-terrain vehicles (ATVs), and other vehicle types. Figure 2 A structural diagram of an electric two-wheeled vehicle provided in this application embodiment is shown below. Figure 2 As shown, the electric two-wheeler 102 integrates four standardized modules: an onboard sensor unit (environmental detection unit), an onboard network transmission unit (communication module), an onboard controller (core logic unit), and an onboard actuator (physical execution unit). It also includes a backup power supply unit, a positioning unit, and a wide-area communication unit. The onboard controller is electrically connected to the onboard sensor unit, onboard network transmission unit, onboard actuator, backup power supply unit, positioning unit, and wide-area communication unit. The onboard network transmission unit establishes an encrypted two-way communication link with the onboard network transmission unit of the wearable security terminal 106 via wireless communication technology. The onboard network transmission unit refers to the module responsible for network communication, including but not limited to 4G modules, Bluetooth modules, WIFI modules, etc., or inherited modules with various network communication capabilities. The onboard sensor unit refers to the environmental detection unit, including but not limited to IMU, radar, camera, vibration sensor, temperature and humidity sensor, etc., used to transmit vehicle / environmental data to the onboard controller via the onboard bus. The vehicle controller refers to the module responsible for core logic control, including but not limited to core controllers such as 4G modules, Bluetooth modules, WIFI modules, SOCs, and MCUs, used to issue control commands to vehicle actuators. Vehicle actuators are modules responsible for executing business logic and performing physical control, including but not limited to vehicle body components such as vibration motors, airbag deployment devices, instruments, ABS, audio systems, horns, and lights. The wide-area communication unit establishes wide-area encrypted communication with the cloud server 104.

[0029] The core functions of the electric two-wheeler 102 include: the onboard sensor unit collects real-time data on vehicle driving status, surrounding environment, and vehicle posture, and transmits it to the onboard controller; the onboard network transmission unit realizes bidirectional encrypted data interaction with the wearable safety terminal 106, receives risk / trigger signals from the wearable safety terminal 106, and sends warning / trigger commands to the wearable safety terminal 106; the onboard controller is responsible for the core logic processing of the whole vehicle: risk identification and classification judgment, bidirectional identity authentication, generation and synchronous issuance of warning / trigger commands, accident data processing, OTA (Over-The-Air) upgrade management, etc.; the onboard actuator executes the onboard controller commands: realizing actions such as onboard airbag deployment, instrument warning display, audible and visual alarms, and vehicle stability control; the positioning unit obtains the precise positioning of the vehicle, and the wide area communication unit realizes data interaction with the cloud, accident reporting, and sending rescue requests, etc.

[0030] Cloud server 104 refers to a remote computing service platform deployed on the Internet. It establishes an encrypted data connection with the electric two-wheeler 102 and a pre-defined channel 108 via a wide-area wireless communication network (such as 4G / 5G) to achieve functions such as device identity three-element binding, accident data storage and analysis, automatic emergency rescue dispatch, remote OTA upgrades, user account management, and cross-platform information synchronization. For example, cloud server 104 may include weather data providers, data providers, advertising data providers, trajectory status data providers, and geographic information system data providers. Figure 3 A structural diagram of a cloud server provided in an embodiment of this application is shown below. Figure 3As shown, cloud server 104 is the central management unit in the cloud, containing a device management module, a user account module, a binding and authentication module, a data storage module, an emergency rescue and dispatch module, and an OTA upgrade module. The device management module registers, monitors, and manages the online status, model version, and operational logs of all wearable terminals and vehicle-mounted devices, enabling visualized management and control throughout the device's lifecycle. The user account module creates independent digital identities for riders, supporting login, information management, and personalized settings, establishing a unique association between users and terminals. The binding and authentication module implements three-element encrypted binding and two-way identity authentication between users, wearable terminals, and vehicles, preventing unauthorized access and signal spoofing, and ensuring communication security. The data storage module encrypts and stores accident data, location information, collision events, and device logs, ensuring data integrity and traceability, supporting accident review and liability determination. The emergency rescue and dispatch module automatically obtains location and user information after an accident is triggered, links with emergency centers and emergency contacts, pushes SOS alerts, and shortens rescue response time. The OTA upgrade module remotely pushes firmware and algorithm updates, supporting distribution, verification, and rollback, continuously optimizing system functions, improving security performance, and extending product lifespan. The cloud server 104 communicates with the electric two-wheeler 102 via wide-area wireless communication technology and with the user's APP (Application) via mobile internet, and can connect to third-party emergency rescue platforms / first aid centers. The core functions of the cloud server 104 include: device identity authentication and three-element binding relationship management, accident data storage and analysis, automatic emergency rescue dispatch, device OTA remote upgrade, and user account management.

[0031] Wearable safety terminal 106 can be traditional mechanical protective devices such as helmets, cycling clothes, and protective gear, as well as smart wearable terminals equipped with information transmission and response functions, such as wired wearable airbags, smart helmets, smart AI glasses, vests, neck protectors, cycling clothes, elbow pads, wrist guards, knee pads, etc. Figure 4 A structural diagram of a wearable security terminal provided in an embodiment of this application is shown below. Figure 4As shown, the wearable safety terminal 106 integrates four standardized modules: a terminal sensor unit, a terminal network transmission unit, a terminal controller, and a terminal actuator unit. It also includes a backup power supply unit and an identification unit. The terminal controller is electrically connected to the terminal sensor unit, the terminal network transmission unit, the terminal actuator unit, and the backup power supply unit. The terminal network transmission unit establishes an encrypted two-way communication link with the onboard network transmission unit of the electric two-wheeled vehicle 102 via wireless communication technology. The identification unit is used to bind and associate with the user's APP and the cloud server 104. The terminal sensor unit collects real-time data on the rider's posture, acceleration, and other status data, and transmits it to the terminal controller. The terminal network transmission unit enables bidirectional encrypted data interaction with the electric two-wheeler 102, receives warning / trigger commands from the electric two-wheeler 102, and sends rider-side risk / trigger signals to the electric two-wheeler 102. The terminal controller autonomously determines the collision / fall risk, receives and verifies the commands from the electric two-wheeler 102, generates warning / deployment commands, and sends them to the terminal actuator. The terminal actuator unit executes the commands from the terminal controller to implement actions such as graded vibration warning and airbag deployment protection. Key event data is stored locally with redundancy to ensure the integrity and traceability of accident data.

[0032] The preset channel 108 can be a mobile terminal, an emergency contact's mobile terminal, or a vehicle management backend rescue center, etc., used to receive information such as device status, accident alarms, and rescue progress pushed by the cloud server 104. The preset channel 108 communicates with the cloud platform via the mobile internet (which falls under the category of wireless communication technology) and completes the binding and association of the wearable safety terminal and the vehicle by scanning the identity identifier. Figure 5 A structural diagram of a mobile terminal provided in an embodiment of this application is shown below. Figure 5As shown, the mobile terminal includes a device binding and activation module, a real-time status viewing module, an accident alarm receiving module, a rescue progress query module, a warning threshold setting module, and an OTA upgrade module. The device binding and activation module is used to complete the three-way binding of the user, wearable device, and vehicle by scanning the QR code of the wearable safety terminal 106 or entering the VIN (Vehicle Identification Number) code of the electric two-wheeler 102, and to activate the safety system, ensuring unique identity authentication and permission initialization. The real-time status viewing module displays the wearable terminal's battery level, communication status, sensor unit (such as vehicle-mounted sensor unit or terminal sensor unit) operating status, and vehicle safety system operating status in real time, providing a visual health monitoring interface. The accident alarm receiving module immediately pushes an APP pop-up and sound alarm when an airbag deployment or high-risk collision event is detected, including the accident time, location, and emergency contact information. The rescue progress query module synchronizes the cloud-based rescue dispatch status, displaying the dispatch of emergency personnel, their estimated arrival time, and on-site feedback, improving user safety and information transparency. The warning threshold setting module allows users to customize collision sensitivity, vibration intensity, and warning levels to adapt to different riding habits and road conditions, enhancing the personalized protection experience. The OTA upgrade module receives firmware and algorithm update packages pushed by the cloud server 104, supporting one-click download, verification, and installation to ensure synchronized iteration and reliable operation of the app and terminal system. The preset channel 108 core functions include: device binding and activation, real-time device status monitoring, customizable warning threshold settings, accident alarm reception, rescue progress query, and OTA upgrade management.

[0033] The aforementioned networks may include, but are not limited to, at least one of the following: wired networks and wireless networks. The aforementioned wired networks may include, but are not limited to, at least one of the following: wide area networks (WANs), metropolitan area networks (MANs), and local area networks (LANs). The aforementioned wireless networks may include, but are not limited to, at least one of the following: Wireless Fidelity (WIFI) and Bluetooth.

[0034] The electric two-wheeled vehicle safety protection method of this application embodiment is implemented by the electric two-wheeled vehicle 102. Figure 6 This is a schematic flowchart of an optional safety protection method for an electric two-wheeled vehicle according to an embodiment of this application, as shown below. Figure 6 As shown, the process of this method may include the following steps:

[0035] Step S602: In response to the electric two-wheeler being powered on, collect the vehicle sensor parameters of the electric two-wheeler and receive environmental data of the electric two-wheeler's environment from the cloud server.

[0036] Step S604: Based on vehicle sensor parameters and environmental data, predict the collision risk of the electric two-wheeler.

[0037] In step S606, in response to the prediction of a collision risk with the electric two-wheeler, at least one safety protection action is performed and a trigger signal is sent to the wearable safety terminal; the at least one safety protection action includes at least one of: audible and visual warning, instrument display and vehicle airbag deployment; the trigger signal is used to instruct the wearable safety terminal to perform at least one safety response action, the at least one safety response action includes: performing a body-sensing warning, wearable airbag deployment and audible and visual warning.

[0038] The electric two-wheeled vehicle safety protection method in this embodiment can be applied to the fields of intelligent transportation and active safety technology, and applied to the safety protection scenario of two-wheeled motorcycle riding.

[0039] Electric two-wheeled vehicles (such as motorcycles) are a mainstream mode of transportation for daily commuting and leisure travel in China. However, their open riding structure results in a much higher rider injury and death rate in accidents compared to four-wheeled vehicles. Currently, safety protection technologies for motorcycle riders in the industry are mainly divided into the following three categories:

[0040] (1) Standalone wearable safety terminal: divided into two types: mechanical pull-cord trigger and electronic autonomous trigger. The mechanical type connects the motorcycle and the wearable safety terminal through a pull cord. When the motorcycle crashes, the pull cord breaks and triggers the wearable safety terminal. The electronic type has a built-in IMU inertial measurement unit, which autonomously detects changes in rider posture and collision acceleration. When the threshold is reached, the wearable safety terminal is triggered.

[0041] (2) Motorcycle onboard passive safety system: High-end large-displacement motorcycles are equipped with onboard IMU, ABS / ESC, millimeter-wave radar / camera and other sensing units. Some models are equipped with onboard airbags in the fuel tank location, which can only protect the rider's chest and abdomen and other local parts in the event of a collision.

[0042] (3) Simple linkage scheme: A few schemes achieve one-way triggering of wearable safety terminal by vehicle sensor unit through wired connection or one-way short-distance communication, without two-way interaction and redundancy protection capabilities.

[0043] However, existing smart devices primarily focus on displaying riding information and recording riding data. Wired airbags rely on a hard connection between the rider and the vehicle to provide safety protection. Existing mechanical or smart wearable devices lack bidirectional information transmission with the vehicle. The need to improve safety and reduce the probability of accidents remains unmet. Specifically, existing wearable safety terminals can only be triggered after a collision / fall, unable to obtain risk signals from the vehicle's sensing units in advance, unable to issue warnings to the rider before a collision occurs, and unable to reduce the probability of accidents. To address the problem of wearable safety terminals' inability to provide early warnings in related technologies, this application uses an electric two-wheeler to collect vehicle sensor parameters and environmental data to predict collision risks. When a collision risk is predicted, safety protection actions are executed and a trigger signal is sent to the wearable safety terminal, achieving pre-collision warning and collaborative protection.

[0044] This application embodiment constructs a full-link collaborative safety architecture consisting of a wearable safety terminal (such as an airbag vest), a two-wheeled vehicle's onboard safety system, a cloud management platform (i.e., a cloud server), and a user's mobile app (i.e., a pre-set channel). The four core units achieve bidirectional data interaction through corresponding communication links. The onboard T-BOX serves as the core relay and control unit on the vehicle side, the wearable safety terminal as the core protection and sensing unit on the rider side, and the cloud platform as the central hub for full-process management and dispatch. A three-tiered communication link—Bluetooth near-field bidirectional communication, CAN bus in-vehicle communication, and 4G / 5G wide-area cloud communication—is employed to achieve full-process safety protection, including pre-event graded early warning, in-event collaborative triggering, and post-event rescue closed-loop.

[0045] Wearable safety terminals and vehicle safety systems use wireless communication technologies (including but not limited to WIFI, P2P, TCP / IP, MQTT, Bluetooth, LoRa, etc.) to achieve encrypted two-way communication, ensuring low latency and high reliability of two-way interaction, and transmitting warning commands, attitude data, and collision trigger signals. The vehicle safety system and in-vehicle perception / actuator use vehicle bus communication (including but not limited to CAN, CAN FD, etc.) to achieve 100% reliable transmission of in-vehicle data, transmitting risk signals, detonation commands, and status information. The vehicle safety system and cloud platform use wide-area wireless communication technologies (including but not limited to 4G / 5G, etc.) for encrypted communication, reporting device status, accident data, and location information, and issuing rescue commands and OTA upgrade packages. The cloud platform and user APP use mobile internet communication to push device status, accident alarms, and rescue progress.

[0046] In some embodiments, existing linkage solutions, which cannot be universally applied, are only compatible with specific brands and models of motorcycles, and cannot be compatible with two-wheeled motorcycles of different displacements and protocols. Furthermore, they lack OTA upgrade capabilities, preventing iterative optimization of algorithms and functions. This application's embodiments are compatible with two-wheeled motorcycles of different brands, displacements, and protocols, requiring no modification to the original vehicle's core structure for installation and adaptation. It covers a wide range of compatible models, solving the problems of poor compatibility, incompatibility with all types of two-wheeled motorcycles, and lack of OTA upgrade capabilities in existing linkage solutions. The universal definition of wireless communication technology allows the system to be compatible with various mainstream wireless connection technologies without being bound to specific communication protocols / hardware. It supports OTA remote upgrades, enabling continuous optimization of warning algorithms, triggering logic, and the addition of new functions, significantly extending the product lifecycle. It can be expanded to integrate with functions such as group riding, insurance claims, and vehicle anti-theft, offering strong scenario scalability.

[0047] Vehicle sensor parameters refer to data collected by onboard sensor units that characterize the vehicle's operating status and surrounding environment. For example, in a pre-emptive graded warning scenario, the onboard sensor unit collects real-time data on vehicle movement and the surrounding environment. These vehicle sensor parameters include the relative distance and speed between the vehicle and obstacles, the vehicle's attitude angle, vehicle acceleration, and vehicle speed. The onboard sensor unit detects environmental risks such as forward collisions, rear-end collisions, sudden pedestrian appearances at intersections, and vehicles approaching from blind spots, or non-immediate collision risks such as vehicle skidding, loss of control, and rollover, and then issues early warnings.

[0048] Environmental data refers to data that is queried and distributed by cloud servers based on vehicle location information, and is used to characterize the environmental conditions in which the vehicle is located. For example, environmental data includes rainfall level, light intensity, road surface condition, visibility, etc.

[0049] Collision risk prediction for electric two-wheelers refers to the process by which the onboard controller of an electric two-wheeler calculates and predicts the probability of a collision based on vehicle sensor parameters and environmental data, and compares the predicted probability with a preset probability threshold. When the predicted probability is greater than the preset probability threshold, the process determines that the electric two-wheeler poses a collision risk. Collision risk refers to high-risk events that may cause personal injury to the rider due to dynamic abnormal conditions such as collisions, abnormal posture, or loss of control during riding.

[0050] In some embodiments, collision risk prediction for electric two-wheelers is performed based on vehicle sensor parameters and environmental data, including: inputting vehicle sensor parameters and environmental data into a pre-trained neural network model, the neural network model outputting a predicted collision probability; and determining that the electric two-wheeler poses a collision risk if the predicted collision probability is greater than a preset probability threshold. The neural network model can employ deep learning models such as convolutional neural networks, recurrent neural networks, and long short-term memory networks, and is trained using historical accident data to output a predicted collision probability based on the input vehicle sensor parameters and environmental data.

[0051] In other embodiments, collision risk prediction for electric two-wheelers is performed based on vehicle sensor parameters and environmental data, including: matching the vehicle sensor parameters and environmental data with a preset risk feature library to determine the risk weight corresponding to each parameter; calculating a comprehensive risk score based on the risk weight corresponding to each parameter; and determining that the electric two-wheeler poses a collision risk if the comprehensive risk score is greater than a preset score threshold. The risk feature library may include different types of risk features and their corresponding risk weights, such as forward collision risk features, rear-end collision risk features, and sideslip risk features.

[0052] When a collision risk is predicted, the electric two-wheeler will execute at least one safety protection action and send a trigger signal to a wearable safety terminal. This at least one safety protection action refers to the electric two-wheeler's onboard controller sending control commands to the onboard actuators via the onboard bus. The onboard actuators then perform physical actions based on these commands, including audible and visual warnings (driving the speaker to emit a sound warning and driving LED lights to emit a light warning), instrument display (displaying warning icons, risk type, and risk level on the dashboard), and airbag deployment (triggering the onboard airbag deployment device to deploy the airbags).

[0053] A trigger signal refers to a control command sent by the onboard controller of an electric two-wheeler to a wearable safety terminal via a near-field communication unit, carrying parameters such as risk type, risk level, and warning method. At least one safety response action refers to the physical action performed by the onboard controller of the wearable safety terminal based on the parameters in the trigger signal after receiving it, including actions such as motion-sensing warnings, wearable airbag deployment, and audible and visual warnings.

[0054] Among them, tactile warning refers to an active warning mechanism that uses physical tactile stimulation applied to the skin or muscles of the human body to convey risk information to riders in a non-visual, non-auditory manner. Its core is the output of low-latency, highly recognizable tactile signals. For example, tactile warning includes driving a vibration motor to perform vibration warning, microcurrent prompts, thermal warnings, and water pressure / hydraulic feedback. Specifically, in the case of a wearable safety terminal that is an airbag vest, tactile warning can use a driven eccentric vibration motor to generate low-frequency intermittent vibrations (such as 5Hz × 0.3s) on the shoulders, back, or chest of the vest to indicate mild risks; or use high-frequency continuous vibration (30Hz × continuous 2s) to achieve forced intervention for severe risks; or apply pulse stimulation through patch-type microcurrent electrodes (1–5mA) to awaken the rider's attention; or use a micro Peltier thermoelectric module to locally heat up (such as raising the back of the neck area by 2–5°C) to create a temperature difference to indicate risks; or use a micro fluid pump to drive the local expansion of the sealed liquid bladder to form a mild pressure prompt.

[0055] Wearable airbag deployment refers to the process where, upon detecting a collision or high-risk posture event, a wearable device autonomously or collaboratively triggers the instantaneous release of gas or chemical propellant via a deployment device. This causes the airbag to inflate within milliseconds, providing physical cushioning protection for the rider's torso. For example, wearable airbag deployment could involve driving a deployment device to trigger the airbag's deployment.

[0056] Audible and visual warnings refer to a multimodal warning mechanism that uses a combination of sound and visible light signals to deliver visual and auditory alerts to riders, enhancing risk perception in noisy or sensory-impaired environments. For example, audible and visual warnings can involve driving a speaker to emit a sound warning or driving LED lights to emit a light warning.

[0057] It should be noted that: at least one safety protection action is performed by the electric two-wheeler, acting on the vehicle end; at least one safety response action is performed by the wearable safety terminal, acting on the rider end. At least one safety response action is triggered by at least one safety protection action, achieving coordinated action through trigger signals to form comprehensive collaborative protection.

[0058] Optionally, after the electric two-wheeler is powered on, the on-board controller initializes each on-board sensor unit, including on-board radar, on-board IMU, and vehicle speed sensor. The on-board controller reads the vehicle sensor parameters collected by each on-board sensor unit through the on-board bus. The on-board controller sends an environmental data request to the cloud server through the wide-area communication unit, and the environmental data request carries the vehicle's current location information. The cloud server queries the corresponding weather data, road condition data, surrounding vehicle data, and other environmental data based on the vehicle location information, and sends them to the electric two-wheeler. The on-board controller receives the environmental data sent by the cloud server. The on-board controller inputs the vehicle sensor parameters and environmental data into a preset collision risk prediction model. The collision risk prediction model is trained based on a machine learning algorithm and can output a predicted collision probability based on the input data. The on-board controller compares the predicted collision probability with a preset probability threshold. When the predicted collision probability is greater than the preset probability threshold, it is determined that the electric two-wheeler has a collision risk. When the predicted collision probability is greater than a preset probability threshold, the vehicle controller generates a safety protection command and sends the control command to the vehicle actuator via the vehicle bus. The vehicle actuator performs at least one safety protection action according to the control command. At the same time, the vehicle controller sends a trigger signal to the wearable safety terminal via the near-field communication unit. After receiving the trigger signal, the wearable safety terminal performs the corresponding safety response action according to the parameters in the trigger signal.

[0059] For example, in a pre-emptive graded warning scenario (two-way linkage, early risk avoidance), the above method is applicable to high-risk situations where, during vehicle riding, the onboard sensor unit detects environmental risks such as forward collisions, rear-end collisions, sudden pedestrian appearances at intersections, and vehicles approaching from blind spots, or non-instantaneous collision risks such as vehicle skidding, loss of control, and rollover. Specifically, the onboard sensor unit collects real-time data on vehicle movement and the surrounding environment. After identifying a risk, it generates a graded warning signal (mild warning / severe warning) and sends it to the onboard controller via the onboard bus. After verifying the warning signal, the onboard controller sends the corresponding graded warning command to the wearable safety terminal in real-time via an encrypted two-way communication link established by wireless communication technology. Simultaneously, it sends warning commands to onboard actuators (such as instruments and audible / visual alarms) via the onboard bus. The onboard terminal executes the warning synchronously; for example, the instrument displays the risk type and warning level, and the audible / visual alarm emits a corresponding frequency of prompt sound and light warning. The wearable safety terminal executes... The warning system works as follows: After receiving a command, the on-board controller drives the on-board actuator (vibration motor) to execute the corresponding vibration mode. For example, a mild warning uses short-interval vibrations to remind the rider of the risk; a severe warning uses high-frequency continuous vibrations to force the rider to take immediate evasive action. Directional warnings can be achieved through on-board actuators in different locations (e.g., vibration is triggered at the corresponding location when a vehicle approaches from the left blind spot). After the risk is cleared, the on-board sensor unit sends a risk clearance signal to the on-board controller. The on-board controller then simultaneously sends a clearance command to the wearable safety terminal via wireless communication technology. The wearable safety terminal and the on-board terminal then stop issuing warnings and resume normal monitoring.

[0060] Through the embodiments provided in this application, electric two-wheeled vehicles predict collision risks by collecting vehicle sensor parameters and environmental data. When a collision risk is predicted, safety protection actions are performed and a trigger signal is sent to a wearable safety terminal. This achieves pre-emptive graded early warning and collaborative protection, solving the problems of existing wearable safety terminals lacking pre-emptive risk warning capabilities, only being able to trigger after the fact, and being unable to avoid accidents in advance. This improves the accuracy of risk identification and the timeliness of protection.

[0061] In one exemplary embodiment, the method further includes: uploading a warning event record to a cloud server in response to a prediction of a collision risk to an electric two-wheeler.

[0062] In one exemplary embodiment, the method further includes: pushing device status, accident alarms, and rescue progress to a preset channel via a cloud server.

[0063] Optionally, when the predicted collision probability exceeds a preset probability threshold, the on-board controller of the electric two-wheeler generates a warning event record. The warning event record includes the warning time, warning location, risk type, risk level, predicted collision probability, vehicle sensor parameter snapshot, and environmental data snapshot. The on-board controller uploads the warning event record to the cloud server via a wide-area communication unit. After receiving the warning event record, the cloud server stores it in the database and returns a confirmation response. The on-board controller receives the confirmation response returned by the cloud server to confirm that the warning event record has been successfully uploaded.

[0064] Optionally, after receiving the warning event record, the cloud server parses the risk level in the warning event record; when the risk level exceeds the preset severe risk threshold, the cloud server determines it as a serious accident and automatically triggers the rescue process; the cloud server queries the user's pre-stored list of emergency contacts and sends an accident notification to the emergency contacts, the accident notification including the accident location, accident time, risk type, risk level, etc.; the cloud server queries the nearest rescue organization based on the accident location and sends a rescue request to the rescue organization; at the same time, the cloud server pushes device status, accident alarms, and rescue progress to the user's mobile application terminal, and the user can view the estimated arrival time of the rescue organization, rescue status, and other information through the mobile application terminal.

[0065] In this embodiment, when an electric two-wheeler anticipates a collision risk, it uploads a warning event record to a cloud server. The cloud server then pushes device status, accident alarms, and rescue progress to preset channels, realizing cloud storage of warning data and an automatic rescue loop after an accident. This facilitates subsequent data analysis, accident tracing, and algorithm optimization, reduces rescue response time, and improves rescue efficiency.

[0066] In one exemplary embodiment, a collision risk prediction for an electric two-wheeler is performed based on vehicle sensor parameters and environmental data, including: determining a first probability of a collision based on the vehicle sensor parameters; determining a second probability of an abnormal posture of the electric two-wheeler based on the vehicle sensor parameters; determining a third probability of a loss of control event due to environmental factors based on the environmental data; fusing the first, second, and third probabilities to obtain a predicted collision probability; and determining that the electric two-wheeler has a collision risk in response to the predicted collision probability being greater than a preset probability threshold.

[0067] The first probability refers to the collision probability calculated based on the relative distance and relative speed between the vehicle and the obstacle. It is used to quantify the risk of a collision between the vehicle and the obstacle. The higher the first probability, the greater the likelihood of a collision.

[0068] In some embodiments, determining the first probability of a collision involving an electric two-wheeled vehicle based on vehicle sensor parameters includes: inputting the vehicle sensor parameters into a pre-trained collision probability prediction model, and the collision probability prediction model outputting the first probability. The collision probability prediction model can employ machine learning models such as support vector machines, random forests, or gradient boosting decision trees, and is trained using historical collision data to output the first probability based on the input vehicle sensor parameters. Vehicle sensor parameters may include the relative distance and speed between the vehicle and obstacles, the vehicle speed, the vehicle acceleration, and the vehicle angular velocity.

[0069] In other embodiments, determining a first probability of a collision involving the electric two-wheeler based on vehicle sensor parameters includes matching the vehicle sensor parameters with a preset collision risk rule base to determine the first probability. The collision risk rule base may include multiple collision risk rules, each corresponding to a probability value. For example, when the relative distance between the vehicle and an obstacle is less than a first distance threshold and the relative speed is less than a first speed threshold, the first probability is 0.9; when the relative distance between the vehicle and an obstacle is less than a second distance threshold and the relative speed is less than a second speed threshold, the first probability is 0.7, and so on.

[0070] The second probability refers to the probability of an abnormal attitude calculated based on the vehicle's attitude angles and acceleration. It is used to quantify the risk level of an abnormal vehicle attitude. The higher the second probability, the greater the likelihood that the vehicle will exhibit abnormal attitudes such as skidding, loss of control, or rollover.

[0071] In some embodiments, determining a second probability of an abnormal posture of an electric two-wheeler based on vehicle sensor parameters includes: inputting the vehicle sensor parameters into a pre-trained posture anomaly detection model, and the posture anomaly detection model outputting the second probability. The posture anomaly detection model can employ deep learning models such as convolutional neural networks, recurrent neural networks, and long short-term memory networks, and is trained using historical posture anomaly data to output the second probability based on the input vehicle sensor parameters. Vehicle sensor parameters may include vehicle posture angles, vehicle acceleration, vehicle angular velocity, and vehicle speed.

[0072] In other embodiments, determining a second probability of an abnormal posture in the electric two-wheeler based on vehicle sensor parameters includes: calculating the Euclidean distance between the vehicle sensor parameters and reference parameters under normal riding conditions, and determining the second probability based on the Euclidean distance. A larger Euclidean distance indicates a greater deviation between the current vehicle state and the normal riding state, and a higher second probability. Reference parameters can be obtained by statistically analyzing a large amount of normal riding data, including average posture angles, average acceleration, and average angular velocity under normal riding conditions.

[0073] The third probability refers to the environmental risk probability calculated based on rainfall level, light intensity, and road conditions. It is used to quantify the risk of a vehicle losing control due to environmental factors. The higher the third probability, the greater the likelihood of a vehicle losing control due to environmental factors.

[0074] In some embodiments, determining a third probability of an electric two-wheeler experiencing a loss of control due to environmental factors, based on environmental data, includes: inputting environmental data into a pre-trained environmental risk assessment model, and the environmental risk assessment model outputting the third probability. The environmental risk assessment model can employ deep learning models such as multilayer perceptron models, deep belief network models, and autoencoder models, and is trained using historical environmental risk data to output the third probability based on the input environmental data. Environmental data may include rainfall levels, light intensity, road surface conditions, visibility, wind speed, and temperature.

[0075] In other embodiments, determining the third probability of an electric two-wheeler losing control due to environmental factors based on environmental data includes: matching the environmental data with a preset environmental risk level table to determine the risk level corresponding to each environmental factor, and calculating the third probability based on the risk level corresponding to each environmental factor. The environmental risk level table may include a rainfall risk level table, a sunlight risk level table, a road surface condition risk level table, etc. For example, the rainfall risk level table may include: light rain corresponding to risk level 1, moderate rain corresponding to risk level 2, heavy rain corresponding to risk level 3, and torrential rain corresponding to risk level 4, etc. A weighted average is calculated as the third probability based on the risk level corresponding to each environmental factor.

[0076] Predicted collision probability refers to the comprehensive risk probability obtained by fusing the first, second, and third probabilities. It is used to quantify the overall risk level of a collision event involving an electric two-wheeler. Predicted collision probability comprehensively considers the collision risk between the vehicle and an obstacle, the risk of abnormal vehicle posture, and the risk of loss of control due to environmental factors, and can comprehensively reflect the collision risk status of electric two-wheelers.

[0077] The preset probability threshold refers to a pre-set probability threshold used to determine whether an electric two-wheeler poses a collision risk. If the predicted collision probability is greater than the preset probability threshold, it means that the overall collision risk of the electric two-wheeler has exceeded the safety warning line, and immediate safety protection measures need to be taken, including executing safety protection actions and sending trigger signals to wearable safety terminals.

[0078] This embodiment calculates the first probability, the second probability, and the third probability separately, and then performs fusion processing to obtain the predicted collision probability. This achieves a comprehensive assessment of multi-dimensional risks, improves the accuracy and comprehensiveness of risk identification, and can more accurately reflect the collision risk status of electric two-wheelers.

[0079] In an exemplary embodiment, the vehicle sensor parameters include the relative distance between the electric two-wheeler and the obstacle, and the relative speed between the electric two-wheeler and the obstacle. Determining a first probability of collision between the electric two-wheeler and the obstacle based on the vehicle sensor parameters includes: determining the first probability of collision between the electric two-wheeler and the obstacle as 0 in response to a relative speed greater than or equal to 0; determining the ratio between the absolute values ​​of the relative distance and the relative speed as the predicted collision duration in response to a relative speed less than 0; determining a target duration range within which the predicted collision duration falls from multiple duration ranges; and determining the probability value corresponding to the target duration range as the first probability of collision between the electric two-wheeler and the obstacle.

[0080] The relative distance and speed between the electric two-wheeler and obstacles are obtained through onboard sensor units. These units may include millimeter-wave radar, lidar, ultrasonic sensors, and cameras. Millimeter-wave radar calculates the relative distance and speed by emitting electromagnetic waves and receiving reflected waves; lidar calculates the relative distance by emitting laser beams and receiving reflected light; ultrasonic sensors calculate the relative distance by emitting ultrasonic waves and receiving reflected waves; and cameras identify obstacles using image processing algorithms and calculate the relative distance and speed based on the obstacle's position in the image. The onboard controller reads the relative distance and speed data collected by each sensor unit via the vehicle bus.

[0081] A relative speed greater than or equal to 0 means that the distance between the electric two-wheeler and the obstacle is increasing or remaining constant, i.e., the electric two-wheeler is moving away from the obstacle or remaining relatively stationary. In this case, the electric two-wheeler will not collide with the obstacle, so the first probability of a collision between the electric two-wheeler and the obstacle must be determined to be 0. This avoids false alarms in situations where there is no risk of collision and reduces interference with the rider.

[0082] A relative speed less than 0 indicates that the distance between the electric two-wheeler and the obstacle is decreasing, meaning the electric two-wheeler is approaching the obstacle. In this situation, there is a risk of collision, and the urgency of the collision needs to be further assessed. The ratio between the absolute values ​​of relative distance and relative speed represents the estimated time required for the electric two-wheeler to collide with the obstacle, i.e., the predicted collision duration. The predicted collision duration refers to the estimated time of collision calculated based on the current relative distance and relative speed, used to quantify the urgency of the collision. The shorter the predicted collision duration, the more imminent the collision, and the higher the collision risk. In this embodiment, the absolute value of the relative speed is used instead of the ratio because a relative speed less than 0 is a negative value, and directly using a negative value for the ratio calculation would result in a negative predicted collision duration, which does not conform to the physical meaning of time; using the absolute value of the relative speed ensures that the predicted collision duration is positive, facilitating subsequent duration range matching and probability value determination.

[0083] Each of the multiple time ranges refers to a pre-defined time interval used to map the predicted collision duration to a corresponding first probability. The division of time ranges can be set according to the timeliness requirements of collision protection and the rider's reaction time. For example, multiple time ranges may include: a first time range (0 to 1 second), a second time range (1 to 3 seconds), a third time range (3 to 5 seconds), a fourth time range (5 to 10 seconds), a fifth time range (more than 10 seconds), etc. Each time range corresponds to a probability value, which represents the probability of the electric two-wheeler colliding with an obstacle. The shorter the time range, the higher the corresponding probability value, indicating a higher collision risk. For example, the first duration range (0 to 1 second) corresponds to a probability value of 1.0, indicating that a collision is imminent and the risk is extremely high; the second duration range (1 to 3 seconds) corresponds to a probability value of 0.8, indicating that a collision is imminent and the risk is very high; the third duration range (3 to 5 seconds) corresponds to a probability value of 0.5, indicating that the risk of a collision is relatively high; the fourth duration range (5 to 10 seconds) corresponds to a probability value of 0.3, indicating that the risk of a collision is moderate; and the fifth duration range (more than 10 seconds) corresponds to a probability value of 0.1, indicating that the risk of a collision is low.

[0084] It should be noted that when the relative speed is less than 0, the predicted collision duration must be determined first, and then the probability value corresponding to the predicted collision duration must be determined by partitioning. This probability value is then set as the first probability of a collision involving an electric two-wheeler because: First, the predicted collision duration can intuitively reflect the urgency of the collision, making it easier for riders to understand and judge; Second, by partitioning the duration range, continuous predicted collision durations can be mapped to discrete probability values, simplifying the risk assessment logic and improving computational efficiency; Third, the division of the duration range can be adjusted according to actual application scenarios, providing strong flexibility and adaptability; Fourth, using the predicted collision duration instead of directly using relative distance and relative speed for risk assessment can more accurately reflect the dynamic changes in collision risk. For example, when the relative distance is large but the relative speed is also large, the predicted collision duration may be short, and the collision risk may be high.

[0085] Optionally, based on vehicle sensor parameters, the first probability of a collision involving the electric two-wheeler is determined, including: the vehicle controller of the electric two-wheeler reads the relative distance D and relative speed V_rel between the vehicle and the obstacle ahead detected by the vehicle radar via the vehicle bus; the vehicle controller determines the sign of the relative speed V_rel; when V_rel≥0, it indicates that the vehicle and the obstacle are moving away from each other or are relatively stationary, and there is no risk of collision, so the vehicle controller determines the first probability P1=0; when V_rel<0, it indicates that the vehicle and the obstacle are approaching each other, and there is a risk of collision, so the vehicle controller calculates the predicted collision duration T=D / |V_rel|; the vehicle controller determines the target duration range in which the predicted collision duration T is located from multiple preset duration ranges; the vehicle controller determines the probability value corresponding to the target duration range as the first probability P1.

[0086] For example, several preset duration ranges and their corresponding probability values ​​are as follows: T < 1 second corresponds to a probability value of 1.0, 1 second ≤ T < 3 seconds corresponds to a probability value of 0.8, 3 seconds ≤ T < 5 seconds corresponds to a probability value of 0.5, 5 seconds ≤ T < 10 seconds corresponds to a probability value of 0.3, and T ≥ 10 seconds corresponds to a probability value of 0.1. When the predicted collision duration T = 2 seconds, the target duration range is 1 second ≤ T < 3 seconds, corresponding to a probability value of 0.8, therefore the first probability P1 = 0.8.

[0087] This embodiment uses the sign of relative speed to determine the existence of collision risk and predicts the collision duration to quantify the urgency of the collision, thus achieving an accurate assessment of collision risk. It can promptly identify high-risk collision events, provide riders with sufficient time to avoid danger, and effectively reduce the accident rate.

[0088] In an exemplary embodiment, the vehicle sensor parameters include the attitude angle and acceleration of the electric two-wheeler; determining a second probability of the electric two-wheeler exhibiting an abnormal attitude based on the vehicle sensor parameters includes: determining the ratio between the absolute value of the attitude angle and an attitude angle threshold as the attitude angle anomaly degree; determining the ratio between the absolute value of the acceleration and an acceleration threshold as the acceleration anomaly degree; and fusing the attitude angle anomaly degree and the acceleration anomaly degree to obtain a second probability of the electric two-wheeler exhibiting an abnormal attitude.

[0089] Attitude angles are used to characterize the vehicle's tilt state. Attitude angles can include at least one of pitch angle, roll angle, and yaw angle. Pitch angle represents the vehicle's tilt angle in the forward / backward direction; a positive value indicates the front of the vehicle is raised, and a negative value indicates the front of the vehicle is lowered. Roll angle represents the vehicle's tilt angle in the left / right direction; a positive value indicates a rightward tilt, and a negative value indicates a leftward tilt. Yaw angle represents the angle between the vehicle's direction of travel and a reference direction. Attitude angles are obtained through an onboard IMU (Inertial Measurement Unit), which integrates a three-axis accelerometer and a three-axis gyroscope. It can measure the vehicle's acceleration and angular velocity in real time and calculate the attitude angles using attitude calculation algorithms (such as complementary filtering, Kalman filtering, etc.). The onboard controller reads the attitude angle data collected by the onboard IMU through the onboard bus.

[0090] Acceleration refers to the motion acceleration of an electric two-wheeled vehicle, used to characterize changes in the vehicle's motion state. Acceleration can include at least one of longitudinal acceleration, lateral acceleration, and vertical acceleration. Longitudinal acceleration represents the acceleration in the forward-backward direction, with positive values ​​indicating acceleration and negative values ​​indicating deceleration; lateral acceleration represents the acceleration in the left-right direction, with positive values ​​indicating rightward deviation and negative values ​​indicating leftward deviation; vertical acceleration represents the acceleration in the up-down direction, with positive values ​​indicating upward motion and negative values ​​indicating downward motion. Acceleration is obtained through an onboard IMU (Integrated Measurement Unit), whose triaxial accelerometers can measure the vehicle's triaxial acceleration in real time. The onboard controller reads the acceleration data collected by the onboard IMU through the onboard bus.

[0091] Attitude angle thresholds are pre-set critical values ​​used to determine whether an attitude angle is abnormal. Attitude angle thresholds can be set according to the range of attitude angles under normal riding conditions; for example, the pitch angle threshold can be set to 15 degrees, and the roll angle threshold can be set to 20 degrees. When the absolute value of an attitude angle exceeds the attitude angle threshold, it indicates an abnormal vehicle attitude angle, which may lead to risks such as sideslip, loss of control, and rollover.

[0092] Acceleration thresholds are pre-set critical values ​​used to determine whether acceleration is abnormal. Acceleration thresholds can be set based on the acceleration range under normal riding conditions; for example, the longitudinal acceleration threshold can be set to 0.5 m / s², and the lateral acceleration threshold can be set to 0.3 m / s². When the absolute value of the acceleration exceeds the acceleration threshold, it indicates abnormal vehicle acceleration, which may lead to risks such as sudden acceleration, sudden deceleration, and sharp turns.

[0093] In this embodiment, the ratio between the absolute value of the attitude angle and the attitude angle threshold is determined as the attitude angle anomaly degree; the ratio between the absolute value of the acceleration and the acceleration threshold is determined as the acceleration anomaly degree. The attitude angle anomaly degree, which is the ratio of the absolute value of the attitude angle to the attitude angle threshold, is used to quantify the degree of anomaly in the vehicle's attitude angle. The attitude angle anomaly degree ranges from 0 to positive infinity; a larger anomaly degree indicates a higher degree of anomaly in the vehicle's attitude angle. When the attitude angle anomaly degree is less than 1, it indicates that the absolute value of the attitude angle is less than the attitude angle threshold, and the attitude angle is within the normal range; when the attitude angle anomaly degree is greater than or equal to 1, it indicates that the absolute value of the attitude angle is greater than or equal to the attitude angle threshold, and the attitude angle is abnormal. The acceleration anomaly degree, which is the ratio of the absolute value of the acceleration to the acceleration threshold, is used to quantify the degree of anomaly in the vehicle's acceleration. The acceleration anomaly degree ranges from 0 to positive infinity; a larger acceleration anomaly degree indicates a higher degree of anomaly in the vehicle's acceleration. When the acceleration anomaly degree is less than 1, it means that the absolute value of the acceleration is less than the acceleration threshold and the acceleration is within the normal range; when the acceleration anomaly degree is greater than or equal to 1, it means that the absolute value of the acceleration is greater than or equal to the acceleration threshold and the acceleration is abnormal.

[0094] It should be noted that the attitude angle anomaly and acceleration anomaly are determined based on the attitude angle and acceleration, and then fused to obtain the second probability. This is because: First, attitude angle and acceleration have different physical meanings and dimensions, and direct fusion would lead to inconsistencies in dimensions. Calculating the anomaly can convert parameters with different dimensions into dimensionless ratios, facilitating fusion. Second, attitude angle and acceleration anomaly can intuitively reflect the degree of deviation of each parameter from the normal state, making it easier to understand the source of risk. Third, by fusing attitude angle and acceleration anomaly, the anomalies in both the attitude angle and acceleration dimensions can be comprehensively considered, improving the accuracy and comprehensiveness of the second probability. Fourth, the fusion process can adjust the weights of attitude angle and acceleration anomalies according to the actual application scenario, exhibiting strong flexibility and adaptability.

[0095] Optionally, the ratio between the absolute value of the attitude angle and the attitude angle threshold is determined as the attitude angle anomaly, including: the on-board controller of the electric two-wheeler extracts the vehicle attitude angle θ from the vehicle sensor parameters; the on-board controller calculates the attitude angle anomaly A_θ=|θ| / θ_max, where θ_max is a preset attitude angle threshold, such as the pitch angle threshold θ_pitch_max=15 degrees, the roll angle threshold θ_roll_max=20 degrees, etc.

[0096] The ratio between the absolute value of acceleration and the acceleration threshold is determined as the acceleration anomaly degree. This includes: the vehicle controller extracting the vehicle acceleration 'a' from the vehicle sensor parameters; the vehicle controller calculating the acceleration anomaly degree A_a = |a| / a_max, where a_max is a preset acceleration threshold, such as the longitudinal acceleration threshold a_long_max = 0.5g, the lateral acceleration threshold a_lat_max = 0.3g, etc.

[0097] The attitude angle anomaly and acceleration anomaly are fused to obtain the second probability of an abnormal attitude in the electric two-wheeler. This includes: the onboard controller calculates the second probability P2 = W_θ × min(A_θ, 1) + W_a × min(A_a, 1), where W_θ and W_a are weighting coefficients, and W_θ + W_a = 1. The maximum value of the anomaly is limited to 1 by the min function to ensure that the second probability P2 is between 0 and 1. For example, when W_θ = 0.6, W_a = 0.4, A_θ = 0.8, and A_a = 1.2, the second probability P2 = 0.6 × 0.8 + 0.4 × 1 = 0.88.

[0098] In this embodiment, by calculating the anomaly degree of attitude angle and the anomaly degree of acceleration separately and then performing fusion processing to obtain the second probability, a comprehensive assessment of vehicle attitude anomalies is achieved. This can accurately identify abnormal attitude risks such as sideslip, loss of control, and rollover, providing riders with timely warnings and protection.

[0099] In one exemplary embodiment, environmental data includes rainfall level, light intensity, and road surface condition.

[0100] Rainfall intensity levels are used to characterize the intensity of rainfall and quantify its impact on cycling safety. Rainfall intensity levels can be obtained through rain sensors or by querying meteorological data from a cloud server. The quantitative value of rainfall intensity levels can be expressed in rainfall amounts (unit: mm / hour), for example, light rain corresponds to 0.1-2.5 mm / hour, moderate rain to 2.6-8.0 mm / hour, heavy rain to 8.1-15.9 mm / hour, and torrential rain to above 16.0 mm / hour. Rainfall intensity levels can also be expressed as numerical levels, for example, level 1 represents light rain, level 2 represents moderate rain, level 3 represents heavy rain, and level 4 represents torrential rain, etc.

[0101] Illumination intensity refers to the intensity of ambient light conditions, used to quantify the impact of light on a cyclist's field of vision. Illumination intensity can be detected by light sensors or obtained from meteorological data retrieved from a cloud server. The quantified value of illuminance can be expressed in lux (unit: lux). For example, sufficient light corresponds to an illuminance of over 10,000 lux, moderate light corresponds to 1,000-10,000 lux, insufficient light corresponds to 100-1,000 lux, and no light corresponds to below 100 lux. Illumination intensity can also be expressed using numerical levels, such as level 1 for sufficient light, level 2 for moderate light, level 3 for insufficient light, and level 4 for no light, etc.

[0102] Road surface condition refers to the state of a road surface, used to quantify its impact on cycling stability. Road surface condition can be detected by road surface sensors or retrieved from cloud servers based on road condition data. Quantitative values ​​for road surface condition can be represented by condition types, such as dry, wet, waterlogged, icy, muddy, and gravel. Road surface condition can also be represented by numerical levels, such as level 1 for dry road surface, level 2 for wet road surface, level 3 for waterlogged road surface, and level 4 for icy road surface.

[0103] In some embodiments, determining a third probability of an electric two-wheeler losing control due to environmental factors based on environmental data includes: determining a target rainfall range containing the rainfall level from multiple rainfall ranges, and determining a target illumination range containing the illumination intensity from multiple illumination ranges; and fusing a first risk probability corresponding to the target rainfall range, a second risk probability corresponding to the target illumination range, and a third risk probability matching the road surface condition to obtain a third probability of an electric two-wheeler losing control due to environmental factors.

[0104] The presence of a risk probability for each of the multiple rainfall ranges indicates that different rainfall intensities have varying degrees of impact on cycling safety. Higher rainfall intensity results in wetter road surfaces, lower tire grip, and a greater risk of loss of vehicle control, thus corresponding to a higher risk probability. For example, light rain (0.1-2.5 mm / hour) corresponds to a risk probability of 0.2, indicating a relatively small impact on cycling safety; moderate rain (2.6-8.0 mm / hour) corresponds to a risk probability of 0.5, indicating a moderate impact; heavy rain (8.1-15.9 mm / hour) corresponds to a risk probability of 0.8, indicating a significant impact; and torrential rain (above 16.0 mm / hour) corresponds to a risk probability of 1.0, indicating an extremely significant impact.

[0105] Each of the multiple illumination ranges corresponds to a risk probability, meaning that different lighting conditions affect the rider's field of vision to varying degrees. The lower the light intensity, the worse the rider's vision, the weaker their perception of the surrounding environment, and the higher the risk of losing control of the vehicle, thus corresponding to a higher risk probability. For example, the risk probability for the well-lit range (above 10,000 lux) is 0.1, indicating that the light has little impact on the rider's field of vision; the risk probability for the moderately lit range (1,000-10,000 lux) is 0.3, indicating that the light has a relatively small impact on the rider's field of vision; the risk probability for the poorly lit range (100-1,000 lux) is 0.6, indicating that the light has a relatively large impact on the rider's field of vision; and the risk probability for the completely dark range (below 100 lux) is 0.9, indicating that the light has a very large impact on the rider's field of vision.

[0106] Different road surface conditions correspond to different risk probabilities, meaning that different road surface conditions have varying degrees of impact on riding stability. The worse the road surface conditions, the lower the coefficient of friction between the tires and the road surface, the worse the vehicle stability, the higher the risk of loss of control, and the greater the corresponding risk probability. For example, the risk probability for a dry road surface is 0.1, indicating that the road surface conditions have a very small impact on riding stability; the risk probability for a wet road surface is 0.4, indicating that the road surface conditions have a moderate impact on riding stability; the risk probability for a flooded road surface is 0.7, indicating that the road surface conditions have a significant impact on riding stability; and the risk probability for an icy road surface is 1.0, indicating that the road surface conditions have a very large impact on riding stability.

[0107] It should be noted that the third probability of an electric two-wheeler losing control due to environmental factors is obtained by fusing the first risk probability corresponding to the target rainfall range, the second risk probability corresponding to the target illumination range, and the third risk probability matched with road conditions. This is because: First, rainfall level, illumination intensity, and road conditions are three main environmental factors affecting riding safety. A single factor cannot fully reflect environmental risk; by fusing the risk probabilities of the three factors, a comprehensive assessment of environmental risk can be achieved. Second, the three environmental factors interact with each other. For example, rain can lead to wet or waterlogged roads, and insufficient illumination can exacerbate the risk of riding in the rain. The fusion process can take into account the interactive effects between factors. Third, the fusion process can adjust the weight of each factor according to the actual application scenario. For example, the weight of illumination intensity can be increased when riding at night, and the weight of rainfall level and road conditions can be increased when riding in the rain, which has strong flexibility and adaptability. Fourth, the third probability obtained through the fusion process can more accurately reflect the risk of loss of control caused by environmental factors, providing riders with more precise warnings.

[0108] Optionally, a target rainfall range for the rainfall level is determined from multiple rainfall ranges, and a target illumination range for the illumination intensity is determined from multiple illumination ranges. This includes: the onboard controller of the electric two-wheeler extracts the rainfall level R and illumination intensity L from environmental data; the onboard controller compares the rainfall level R with multiple preset rainfall ranges to determine the target rainfall range for the rainfall level R; and the onboard controller compares the illumination intensity L with multiple preset illumination ranges to determine the target illumination range for the illumination intensity L.

[0109] Optionally, the first risk probability corresponding to the target rainfall range, the second risk probability corresponding to the target illumination range, and the third risk probability matching the road surface condition are fused to obtain the third probability of an electric two-wheeler losing control due to environmental factors. This includes: the on-board controller extracting the road surface condition S from the environmental data; the on-board controller querying the third risk probability P_S matching the road surface condition S from a preset road surface condition risk mapping table; the on-board controller querying the first risk probability P_R corresponding to the target rainfall range from a preset rainfall range risk mapping table; the on-board controller querying the second risk probability P_L corresponding to the target illumination range from a preset illumination range risk mapping table; and the on-board controller calculating the third probability P3 = W_R × P_R + W_L × P_L + W_S × P_S, where W_R, W_L, and W_S are weighting coefficients, and W_R + W_L + W_S = 1.

[0110] For example, when the rainfall level R is moderate rain (first risk probability P_R=0.5), the light intensity L is insufficient light (second risk probability P_L=0.6), the road surface condition S is wet road surface (third risk probability P_S=0.4), and the weighting coefficients W_R=0.4, W_L=0.3, and W_S=0.3, the third probability P3=0.4×0.5+0.3×0.6+0.3×0.4=0.5.

[0111] This embodiment calculates the risk probabilities corresponding to rainfall level, light intensity, and road surface condition, and then performs fusion processing to obtain a third probability, achieving a comprehensive assessment of environmental risks. It can accurately identify the risk of loss of control caused by environmental factors and provide riders with targeted warnings and protection suggestions.

[0112] In an exemplary embodiment, the wearable safety terminal and the vehicle safety system are independent of each other and rely on a single data source for collision risk determination, which has the following drawbacks: (1) The vehicle sensor unit cannot sense the rider's real-time status. When the rider's posture is abnormal (such as excessive body tilt, sudden shaking, etc.) but the vehicle's movement is normal, the vehicle sensor unit cannot identify the collision risk, resulting in missed detection; (2) The wearable safety terminal can only detect the rider's own status and cannot obtain information about the vehicle's surrounding environment. When the vehicle faces external collision risk but the rider's posture is normal, the wearable safety terminal also cannot identify the risk; (3) The failure of a single data source will lead to protection failure. For example, when the vehicle sensor unit fails or the wearable safety terminal communication is interrupted, the collision risk detection function is completely lost, and redundancy backup cannot be formed, resulting in poor protection reliability.

[0113] To address the aforementioned shortcomings, this embodiment integrates vehicle sensor parameters (used to calculate the predicted collision probability) and rider-side wearable sensor parameters (used to calculate the fourth probability), enabling data collaboration between the vehicle terminal and the wearable safety terminal. When any data source detects an anomaly, a risk warning can be triggered, forming a redundant backup mechanism. This avoids the protection failure caused by the failure of a single vehicle sensor unit, improving the accuracy and comprehensiveness of collision risk identification.

[0114] In some embodiments, the method further includes: receiving wearable sensor parameters from the wearable safety terminal via near-field communication, and determining a fourth probability of a collision between the wearable safety terminal and the target object based on the wearable sensor parameters; and fusing the fourth probability and the predicted collision probability to obtain a new predicted collision probability of a collision between the electric two-wheeled vehicle and the target object.

[0115] Wearable sensor parameters refer to data collected by wearable safety terminals to characterize the user's state. These parameters can include physiological and kinematic parameters such as user posture angle, acceleration, angular velocity, heart rate, body temperature, and blood pressure. The vehicle-mounted sensor units integrated into the wearable safety terminal can include an IMU (Inertial Measurement Unit), heart rate sensor, body temperature sensor, and blood pressure sensor. The IMU can collect real-time motion data such as user posture angle, acceleration, and angular velocity; the heart rate sensor can collect real-time heart rate data; the body temperature sensor can collect real-time body temperature data; and the blood pressure sensor can collect real-time blood pressure data. The vehicle-mounted controller of the wearable safety terminal transmits the data collected by the vehicle-mounted sensor units to the electric two-wheeler via a near-field communication unit.

[0116] The target audience refers to the users of the wearable safety terminal, namely, the riders of electric two-wheeled vehicles. The target audience's status data is collected through the wearable safety terminal, which can reflect the rider's real-time movement and physiological status.

[0117] The fourth probability refers to the probability of abnormal user posture or user collision risk calculated based on wearable sensor parameters. It is used to quantify the degree of risk of a collision with a target object. The fourth probability comprehensively considers parameters such as the user's posture angle, acceleration, angular velocity, and heart rate, and can reflect the collision risk on the rider's side. The higher the fourth probability, the greater the likelihood of a collision with the target object.

[0118] In some embodiments, determining a fourth probability of a collision between a wearable security terminal and a target object based on wearable sensor parameters includes: inputting the wearable sensor parameters into a pre-trained user posture anomaly detection model, and the user posture anomaly detection model outputting the fourth probability. The user posture anomaly detection model can employ deep learning models such as convolutional neural networks, recurrent neural networks, and long short-term memory networks, and is trained using historical user posture anomaly data to output the fourth probability based on the input wearable sensor parameters. Wearable sensor parameters may include user posture angle, user acceleration, user angular velocity, user heart rate, etc.

[0119] In other embodiments, determining a fourth probability of collision between the wearable safety terminal and the target object based on wearable sensor parameters includes: calculating the Mahalanobis distance between the wearable sensor parameters and baseline parameters under normal riding conditions, and determining the fourth probability based on the Mahalanobis distance. Mahalanobis distance takes into account the correlation between parameters and is more suitable for multi-parameter distance calculations than Euclidean distance. Baseline parameters can be obtained by statistically analyzing a large amount of normal riding data, including the average user posture angle, average user acceleration, average user angular velocity, and average user heart rate under normal riding conditions. A larger Mahalanobis distance indicates a greater deviation between the current user state and the normal riding state, resulting in a higher fourth probability.

[0120] In other embodiments, a fourth probability of collision between the wearable safety terminal and the target object is determined based on wearable sensor parameters, including: the vehicle controller extracting motion state parameters such as user attitude angle θ_u, user acceleration a_u, and user angular velocity ω_u from the wearable sensor parameters; the vehicle controller calculating the user attitude angle anomaly A_θu=|θ_u| / θ_u_max, where θ_u_max is the user attitude angle threshold; and the vehicle controller calculating the user acceleration anomaly A_au=|a_u| / a_u_max, where a_u_max is the user's acceleration threshold. Acceleration threshold; The vehicle controller calculates the user angular velocity anomaly A_ωu=|ω_u| / ω_u_max, where ω_u_max is the user angular velocity threshold; The vehicle controller weights and fuses the user attitude angle anomaly, user acceleration anomaly and user angular velocity anomaly to obtain the fourth probability P4=W_θu×min(A_θu,1)+W_au×min(A_au,1)+W_ωu×min(A_ωu,1), where W_θu, W_au, and W_ωu are weight coefficients, and W_θu+W_au+W_ωu=1.

[0121] The new predicted collision probability is a comprehensive collision risk probability obtained by fusing the fourth probability and the original predicted collision probability. It is used to quantify the overall risk level of a collision event involving an electric two-wheeler. The new predicted collision probability comprehensively considers both the collision risk on the vehicle side (calculated from vehicle sensor parameters) and the collision risk on the rider side (calculated from wearable sensor parameters), thus providing a more comprehensive reflection of the collision risk status of the electric two-wheeler. The new predicted collision probability is applied to subsequent risk assessment and safety protection processes. When the new predicted collision probability exceeds a preset probability threshold, a collision risk is determined for the electric two-wheeler, triggering safety protection actions and sending a trigger signal.

[0122] In some embodiments, the fourth probability and the predicted collision probability are fused to obtain a new predicted collision probability for an electric two-wheeler collision. This includes: the onboard controller calculating the new predicted collision probability P_new = α × P_old + (1-α) × P4, where P_old is the original predicted collision probability, and α is the fusion coefficient, ranging from 0 to 1. The fusion coefficient α can be set according to the reliability of the vehicle sensor parameters and the wearable sensor parameters. For example, when the reliability of the vehicle sensor parameters is high, α takes a larger value; when the reliability of the wearable sensor parameters is high, α takes a smaller value.

[0123] For example, when the original predicted collision probability P_old=0.6, the fourth probability P4=0.8, and the fusion coefficient α=0.7, the new predicted collision probability P_new=0.7×0.6+0.3×0.8=0.66.

[0124] This embodiment integrates wearable sensor parameters to achieve data collaboration between the vehicle terminal and the wearable safety terminal. It can comprehensively consider the collision risks on both the vehicle side and the rider side, improving the accuracy and comprehensiveness of risk identification and providing riders with more precise warnings and protection.

[0125] In one exemplary embodiment, the method further includes: establishing a near-field communication connection with a wearable security terminal via a wireless transmission protocol; receiving wearable sensor parameters periodically transmitted by the wearable security terminal, the wearable sensor parameters including at least one of user physiological data and user operation instructions; and performing security protection decisions based on the wearable sensor parameters.

[0126] User physiological data refers to data used to characterize a user's physiological state, reflecting their health status and physiological changes. User physiological data may include heart rate, body temperature, blood pressure, and respiratory rate. This data is collected through physiological sensors integrated into wearable safety terminals, such as heart rate sensors, body temperature sensors, and blood pressure sensors.

[0127] User operation commands refer to control instructions input by the user through the wearable security terminal, used to trigger specific security protection functions or services. User operation commands can include emergency assistance commands, function switching commands, warning confirmation commands, and warning cancellation commands. Emergency assistance commands are user-initiated distress signals used to request emergency rescue; function switching commands are user-initiated function switching signals used to switch the wearable security terminal's operating mode, such as from normal mode to sports mode or silent mode; warning confirmation commands are user confirmation feedback of warning signals, used to confirm that the user has received the warning and taken appropriate measures; warning cancellation commands are user-initiated commands to cancel warning signals, used to cancel false alarms or processed warnings. User operation commands are input through input units integrated into the wearable security terminal, such as physical buttons, touchscreens, voice recognition modules, and gesture recognition modules.

[0128] Safety protection decisions refer to safety protection strategies or actions generated based on wearable sensor parameters to ensure user safety. Safety protection decisions can include health warning decisions, emergency rescue decisions, function adjustment decisions, and warning escalation decisions. Health warning decisions generate health warning instructions and provide health tips or suggestions to users when their physiological data is abnormal. Emergency rescue decisions generate emergency rescue requests, upload them to the cloud server, and notify emergency contacts when users trigger emergency help commands or detect serious abnormalities. Function adjustment decisions adjust the function settings of wearable safety terminals or electric two-wheelers based on user operation instructions or user status, such as adjusting warning sensitivity or switching working modes. Warning escalation decisions upgrade the warning level or warning method based on user feedback or the duration of the warning, such as escalating from a mild warning to a severe warning.

[0129] In some embodiments, performing safety protection decisions based on wearable sensor parameters includes: parsing the wearable sensor parameters and extracting user physiological data; comparing the user physiological data with a preset normal range of physiological data; generating a health warning command when the user's physiological data exceeds the normal range; sending the health warning command to a wearable security terminal, instructing the wearable security terminal to output health reminder information; and simultaneously pushing a health warning notification to the user's mobile application terminal to remind the user to pay attention to their health status. For example, when the user's heart rate exceeds 120 beats per minute, a health warning command for high heart rate is generated, a voice prompt of "Heart rate too high, please rest" is issued to the user, and a heart rate abnormality notification is pushed to the user's mobile application terminal.

[0130] In other embodiments, safety protection decisions are made based on wearable sensor parameters, including: parsing the wearable sensor parameters and extracting user operation commands; and executing corresponding safety protection actions based on the type of user operation command. When the user operation command is an emergency help command, an emergency rescue request is immediately uploaded to the cloud server. The rescue request includes vehicle location, user information, and the time of request. After receiving the rescue request, the cloud server sends a help notification to the user's emergency contact, sends a rescue request to the nearest rescue organization, and pushes the rescue progress to the user's mobile application terminal. When the user operation command is a function switching command, the function settings of the wearable safety terminal or electric two-wheeler are adjusted, such as adjusting the warning sensitivity or switching the working mode, and the function switching result is fed back to the user. When the user operation command is a warning confirmation command, the user's confirmation feedback is recorded for subsequent warning algorithm optimization. When the user operation command is a warning cancellation command, the current warning action is stopped, normal monitoring is restored, and the warning cancellation event is recorded.

[0131] This embodiment establishes a near-field communication connection, enabling real-time data interaction between the electric two-wheeler and the wearable safety terminal; by receiving wearable sensor parameters, real-time monitoring of the user's status is achieved; and safety protection decisions are executed based on the wearable sensor parameters, realizing personalized safety protection. It can respond promptly to users' health abnormalities and emergency assistance needs, thereby improving the user's safety level.

[0132] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided. Optionally, in this embodiment, the above-described safety protection method for an electric two-wheeled vehicle may be applied, but is not limited to, to applications such as... Figure 1 The hardware environment shown includes an electric two-wheeler 102, a cloud server 104, a wearable security terminal 106, and a preset channel 108.

[0133] The electric two-wheeled vehicle safety protection method of this application embodiment is implemented by the electric two-wheeled vehicle 102. Figure 7 This is a schematic flowchart of an optional safety protection method for an electric two-wheeled vehicle according to an embodiment of this application, as shown below. Figure 7 As shown, the process of this method may include the following steps:

[0134] In step S702, in response to the electric two-wheeler being powered on, the collision risk of the electric two-wheeler is continuously predicted, and communication is made with the wearable safety terminal through an encrypted two-way communication link.

[0135] Step S704: In response to the prediction of a collision risk with the electric two-wheeler, at least one safety protection action is executed, and a trigger signal is sent to the wearable safety terminal; the at least one safety protection action includes at least one of: audible and visual warning, instrument display, and vehicle airbag deployment; the trigger signal is used to instruct the wearable safety terminal to execute at least one safety response action, the at least one safety response action includes: executing a motion-sensing warning, wearable airbag deployment, or audible and visual warning;

[0136] Step S706: In response to receiving a synchronization trigger signal sent by the wearable safety terminal, at least one safety protection action is performed; the synchronization trigger signal indicates that the wearable safety terminal predicts that there is a collision risk with the electric two-wheeler.

[0137] The electric two-wheeled vehicle safety protection method in this embodiment can be applied to the field of safety protection for two-wheeled vehicles such as two-wheeled motorcycles and electric two-wheeled vehicles, and can be applied to scenarios where riders monitor collision risks in real time and trigger safety protection during riding.

[0138] Currently, riders primarily use independent wearable safety terminals for safety protection. However, existing wearable safety terminals and vehicle-mounted safety systems operate independently. Triggering of the wearable safety terminal relies solely on its own judgment; the vehicle passively receives the "trigger request" or has no interaction whatsoever, making timing synchronization impossible. During a collision, the triggering timing of the wearable safety terminal and the vehicle-mounted safety system may misalign, failing to provide comprehensive collaborative protection. Furthermore, it only supports unidirectional triggering without redundancy; the failure of a single vehicle-mounted sensor unit will result in protection failure. Even in systems that support vehicle-triggered wearable safety terminals, it is still only unidirectional. When vehicle-mounted sensor units fail due to obstruction, malfunction, or signal delay, timely triggering of the wearable safety terminal cannot be guaranteed.

[0139] To address the technical problem of unreliable protection system responses caused by the reliance on single-point judgment for triggering of wearable safety terminals in related technologies, this embodiment constructs a bidirectional redundant triggering mechanism between the electric two-wheeler and the wearable safety terminal. This enables both the electric two-wheeler and the wearable safety terminal to have independent collision risk prediction capabilities and safety protection triggering capabilities. If either party detects a collision risk, it can trigger the other party to perform safety protection actions, forming a dual guarantee and avoiding the protection failure problem caused by the failure of a single vehicle-mounted sensor unit.

[0140] This embodiment addresses the problems of existing wearable safety terminals and electric two-wheelers being unable to link bidirectionally, having asynchronous triggering sequences, being prone to failure due to a single triggering source, and having poor protective effects. Through a bidirectional redundant synchronous triggering mechanism, the triggering latency of the wearable safety terminal is significantly lower than the industry standard, greatly improving the protective effect of the wearable safety terminal; bidirectional redundant triggering increases triggering reliability by 100% and significantly reduces the false triggering rate; the collaborative protection between the wearable safety terminal and the vehicle safety system can significantly reduce the rate of serious injury to riders in accidents.

[0141] In this embodiment, both the wearable safety terminal and the electric two-wheeler adopt a standardized modular architecture of "sensor-network transmission unit-controller-actuator". Through general wireless communication technology, bidirectional redundant triggering is achieved, which supports both the wearable terminal to actively trigger vehicle protection and the vehicle to actively trigger wearable protection. This solves the problems of easy failure of single trigger source, asynchronous triggering sequence and limited communication technology binding in the existing technology.

[0142] In some embodiments, this embodiment designs a graded early warning linkage mechanism for typical risk scenarios of electric two-wheeled vehicles. The on-board sensor unit identifies risks in advance and drives the wearable safety terminal to execute graded early warnings through wireless communication technology, which solves the problem that existing wearable safety terminals can only trigger after the fact and have no early warning capability.

[0143] Among them, collision risk prediction for electric two-wheelers refers to the process by which the on-board controller of an electric two-wheeler calculates and predicts the collision probability based on vehicle sensor parameters and environmental data, compares the predicted collision probability with a preset probability threshold, and determines that there is a collision risk for the electric two-wheeler when the predicted collision probability is greater than the preset probability threshold.

[0144] In some embodiments, the electric two-wheeler continuously performs collision risk prediction, including:

[0145] The system acquires vehicle driving status parameters and surrounding environment perception data of the electric two-wheeler; inputs the vehicle driving status parameters and surrounding environment perception data into a pre-trained collision risk prediction model, and the collision risk prediction model outputs the predicted collision probability; in response to the predicted collision probability being greater than a preset probability threshold, it determines that the electric two-wheeler has a collision risk.

[0146] The vehicle driving state parameters include at least one of vehicle speed, acceleration, attitude angle, and steering angle. The surrounding environment perception data includes at least one of the following: obstacle distance, relative speed, and azimuth angle detected by millimeter-wave radar; lane line position, traffic sign information, and type of vehicle ahead detected by cameras. The collision risk prediction model employs a long short-term memory network model, trained using historical collision event data, and can output a predicted collision probability based on the input vehicle driving state parameters and surrounding environment perception data.

[0147] In other embodiments, the electric two-wheeler continuously performs collision risk prediction, including:

[0148] The system acquires the vehicle driving status parameters of the electric two-wheeler and receives environmental risk factors from the cloud server; it calculates the vehicle side collision risk index based on the vehicle driving status parameters and the environmental side collision risk index based on the environmental risk factors; it performs weighted fusion of the vehicle side collision risk index and the environmental side collision risk index to obtain the predicted collision risk index; and it determines that the electric two-wheeler has a collision risk if the predicted collision risk index is greater than a preset risk threshold.

[0149] The environmental risk factor is obtained by the cloud server from meteorological databases, road condition databases, and accident-prone area databases based on the current location of the electric two-wheeler. It includes at least one of the following: weather risk factor, road condition risk factor, and accident-prone area risk factor. The vehicle side-collision risk index is calculated based on parameters such as vehicle speed, acceleration, and attitude angle, reflecting the collision risk posed by the vehicle's own condition. The environmental side-collision risk index is calculated based on the environmental risk factor, reflecting the collision risk posed by the external environment. During weighted fusion, the weights of the vehicle side-collision risk index and the environmental side-collision risk index are determined based on historical data statistical analysis; for example, the weight of the vehicle side-collision risk index is 0.7, and the weight of the environmental side-collision risk index is 0.3.

[0150] When a collision risk is predicted, the electric two-wheeler executes at least one safety protection action and sends a trigger signal to a wearable safety terminal to instruct the wearable safety terminal to execute at least one safety response action. The at least one safety protection action and at least one safety response action have been described in the above embodiments and will not be repeated here.

[0151] While continuously predicting collision risks for the electric two-wheeler, it also communicates with the wearable safety terminal through an encrypted two-way communication link. During this process, the wearable safety terminal continuously predicts collision risks for the electric two-wheeler. When the wearable safety terminal predicts that there is a collision risk for the electric two-wheeler, it executes at least one safety response action and sends a synchronous trigger signal to the electric two-wheeler to instruct it to perform at least one safety protection action.

[0152] The synchronization trigger signal refers to the control command generated and sent to the electric two-wheeler by the wearable safety terminal after detecting a collision risk. The synchronization trigger signal carries at least one of the following: the device identifier of the wearable safety terminal, the collision risk type, the predicted collision probability, the trigger timestamp, and a summary of wearable sensor parameters. The synchronization trigger signal is used to instruct the electric two-wheeler to perform safety protection actions, realizing a redundant backup mechanism for the wearable safety terminal to actively trigger vehicle-side safety protection.

[0153] In some embodiments, the wearable safety terminal continuously predicts the collision risk of electric two-wheelers, including:

[0154] The wearable safety terminal collects motion data such as rider posture angle, acceleration, and angular velocity through a built-in inertial measurement unit; it calculates the posture deviation angle between the rider's posture angle and the normal riding posture angle reference value, and calculates the acceleration deviation value between the rider's acceleration and the normal riding acceleration reference value; when the posture deviation angle is greater than a preset posture deviation threshold, or the acceleration deviation value is greater than a preset acceleration deviation threshold, it determines that the rider's posture is abnormal and generates a posture abnormality score; it calculates the predicted collision probability based on the posture abnormality score, and when the predicted collision probability is greater than a preset probability threshold, it determines that the electric two-wheeler has a collision risk.

[0155] The baseline values ​​for normal riding posture angle and normal riding acceleration are obtained by statistically analyzing a large amount of normal riding data, reflecting the rider's average posture angle and average acceleration under normal riding conditions. The abnormal posture score is calculated by weighting the posture deviation angle and acceleration deviation value. The larger the posture deviation angle and the larger the acceleration deviation value, the higher the abnormal posture score, indicating that the rider's posture is more abnormal and the risk of collision is higher.

[0156] In other embodiments, the wearable safety terminal continuously predicts the collision risk of the electric two-wheeler, including:

[0157] The wearable safety terminal collects the rider's posture angle sequence, acceleration sequence, and angular velocity sequence within a preset time window through a built-in inertial measurement unit; the posture angle sequence, acceleration sequence, and angular velocity sequence are input into a pre-trained rider posture anomaly detection model, and the rider posture anomaly detection model outputs the posture anomaly probability; in response to the posture anomaly probability being greater than a preset anomaly probability threshold, the rider's posture is determined to be abnormal, and thus the electric two-wheeler is determined to have a collision risk.

[0158] The rider posture anomaly detection model employs a hybrid deep learning model combining convolutional neural networks (CNNs) and long short-term memory (LSTM) networks. The CNNs extract local features from posture angle sequences, acceleration sequences, and angular velocity sequences, while the LSTM networks capture long-term dependencies in the time series. The model is trained using historical rider posture anomaly event data and can output the probability of posture anomalies based on the input time series data. The preset anomaly probability threshold is determined based on statistical analysis of historical data, for example, set to 0.7.

[0159] In this embodiment, both the electric two-wheeler and the wearable safety terminal continuously predict the collision risk of the electric two-wheeler. If either predicts a collision risk, it notifies the other and instructs it to respond. This forms a bidirectional redundant triggering mechanism. When the on-board sensor unit fails due to obstruction, malfunction, or signal delay, the wearable safety terminal can still independently detect the collision risk and trigger vehicle-side safety protection. When the wearable safety terminal fails due to communication interruption or on-board sensor unit failure, the vehicle side can still independently detect the collision risk and trigger the wearable safety terminal to execute a safety response. This avoids the protection failure problem caused by the failure of a single on-board sensor unit and improves the reliability and robustness of the safety protection system.

[0160] Figure 8 An optional bidirectional redundant branch is provided for embodiments of this application, such as Figure 8 As shown, this embodiment sets up two bidirectional redundant branches for active triggering by the wearable safety terminal and active triggering by the electric two-wheeler, to avoid protection failure caused by the failure of a single vehicle-mounted sensor unit, and to ensure low latency and high reliability of the wearable safety terminal triggering.

[0161] Branch 1: Wearable safety terminal active triggering branch. This branch is applicable when the wearable safety terminal's terminal sensor unit detects a collision acceleration or attitude change exceeding a threshold before the electric two-wheeler's onboard sensor unit. Examples include situations where the electric two-wheeler overturns, crashes, or the rider is thrown from the vehicle by a collision. The complete interaction and linkage process of the wearable safety terminal active triggering branch includes the following steps:

[0162] In step S801, the terminal sensor unit of the wearable safety terminal collects the rider's motion posture and acceleration data in real time.

[0163] Step S802: Determine whether the detected acceleration and angular velocity exceed the preset collision threshold. If the acceleration and angular velocity exceed the preset collision threshold, proceed to step S803; if the acceleration and angular velocity do not exceed the preset collision threshold, continue to step S801.

[0164] In step S803, the wearable security terminal immediately determines that it is a high-risk collision event and executes step S804.

[0165] In step S804, the terminal controller of the wearable safety terminal synchronously performs two actions: driving the terminal actuator to trigger the wearable safety terminal to perform at least one safety response action (such as the deployment of a wearable airbag); and sending a synchronous trigger signal (such as a collision deployment event signal) to the electric two-wheeler in real time through an encrypted two-way communication link.

[0166] In step S805, the on-board controller of the electric two-wheeler receives the signal and verifies the synchronization trigger signal. If the verification is successful, proceed to step S808; if the verification fails, the process ends.

[0167] In step S808, the electric two-wheeler is immediately identified as a high-risk collision event.

[0168] In step S809, the electric two-wheeler immediately sends a synchronous detonation command to the on-board actuator via the vehicle bus, triggering at least one safety protection action of the on-board actuator (such as the deployment of the on-board airbag), thereby realizing the coordinated synchronous triggering of the wearable safety terminal and the electric two-wheeler.

[0169] In step S810, simultaneously, the on-board controller of the electric two-wheeler obtains the precise location of the accident through the positioning unit, and reports the accident data, positioning information, and rider information to the cloud server through wide-area wireless communication technology, triggering an emergency rescue request.

[0170] In step S811, after receiving the rescue request, the cloud server will issue an alarm and automatically push the accident information and location to the user's emergency contacts, the nearest emergency center or rescue platform, and at the same time push the accident alarm and rescue progress to the user's APP.

[0171] If communication is interrupted, the wearable safety terminal and the electric two-wheeler store the entire collision process data locally, and automatically re-transmit it to the cloud server after communication is restored to ensure that the accident data is complete and traceable.

[0172] Branch Two: Active Triggering Branch for Electric Two-Wheeled Vehicles. This branch is applicable to situations where the onboard sensor unit detects a collision event before the wearable safety terminal. For example, in the event of a frontal collision or rear-end collision involving an electric two-wheeled vehicle, coordinated protection is triggered in advance. The complete interaction and linkage process of the active triggering branch for electric two-wheeled vehicles includes the following steps:

[0173] In step S806, the on-board sensor unit of the electric two-wheeler collects vehicle sensing parameters in real time.

[0174] Step S807: Based on the vehicle sensing parameters, determine whether the detected vehicle collision acceleration or attitude change exceeds a preset collision threshold. If it exceeds the preset collision threshold, proceed to step S808. If it does not exceed the preset collision threshold, continue to step S806.

[0175] In step S808, the electric two-wheeler immediately determines that a collision event has occurred and sends a collision trigger signal to the vehicle controller of the electric two-wheeler via the vehicle bus.

[0176] In step S809, after the on-board controller of the electric two-wheeler receives the signal and completes the verification, it sends a trigger command to the on-board actuator through the on-board bus to trigger at least one safety protection action of the on-board actuator (such as the deployment of the on-board airbag).

[0177] In step S812, the electric two-wheeler sends a trigger signal (such as a collision point explosion event signal) to the wearable safety terminal in real time via an encrypted two-way communication link.

[0178] In step S813, the terminal controller of the wearable security terminal receives the trigger signal and verifies it. If the verification is successful, step S803 is executed; if the verification fails, the process ends.

[0179] The subsequent emergency rescue closed-loop process of Branch 2 is completely consistent with that of Branch 1, and simultaneously triggers actions such as cloud-based emergency rescue, location reporting, and notification of emergency contacts.

[0180] For example, when an electric two-wheeler encounters a vehicle suddenly changing lanes from the side, the millimeter-wave radar on the vehicle's side detects the rapidly decreasing distance to the obstacle and calculates a predicted collision probability of 0.85, which is greater than the preset probability threshold of 0.6. The electric two-wheeler determines there is a collision risk, executes an audible and visual warning, and sends a trigger signal to a wearable safety terminal, instructing the terminal to execute a motion-sensing warning. Simultaneously, if the rider is startled and suddenly leans to the other side, the wearable safety terminal detects this sudden change in rider posture angle, calculates an abnormal posture score of 0.9, determines there is a collision risk, executes a motion-sensing warning, and sends a synchronized trigger signal to the electric two-wheeler, instructing it to execute an audible and visual warning. Upon receiving the synchronized trigger signal, the electric two-wheeler confirms that both the vehicle's side and the wearable safety terminal have detected a collision risk and continues to execute the audible and visual warning, forming a dual-warning protection system.

[0181] For example, when an electric two-wheeler is traveling at night, the camera on the side of the vehicle cannot effectively identify obstacles ahead due to insufficient light, and the millimeter-wave radar fails to detect the collision risk because the obstacle is in a blind spot. The vehicle-side calculated predicted collision probability of 0.3, which is less than the preset probability threshold of 0.6, and therefore no collision risk is detected. However, if the rider brakes suddenly upon seeing an obstacle ahead, causing a sudden increase in the rider's forward lean angle, the wearable safety terminal detects the change in the rider's posture angle, calculates an abnormal posture score of 0.8, determines that there is a collision risk with the electric two-wheeler, executes a motion-sensing warning action, and sends a synchronous trigger signal to the electric two-wheeler. Upon receiving the synchronous trigger signal, the electric two-wheeler executes an audible and visual warning action, avoiding missed warnings caused by the failure of the onboard sensor unit.

[0182] Optionally, after the electric two-wheeler is powered on, the onboard controller initializes the onboard sensor unit, communication unit, and execution unit, and establishes a communication connection with the wearable safety terminal through an encrypted two-way communication link. The onboard controller continuously collects vehicle sensor parameters, including vehicle speed, acceleration, attitude angle, obstacle distance, and relative obstacle speed, and receives environmental data from the cloud server, including weather information, road condition information, and accident-prone area information. The onboard controller calculates the predicted collision probability based on the vehicle sensor parameters and environmental data. Specifically, it calculates the vehicle-side collision risk index based on the vehicle sensor parameters and the environmental-side collision risk index based on the environmental data, and then performs a weighted fusion of the two to obtain the predicted collision probability. The onboard controller determines whether the predicted collision probability is greater than a preset probability threshold. If not, it returns to continue monitoring; if so, it determines that the electric two-wheeler has a collision risk, and the onboard controller executes at least one safety protection action, including controlling the audible and visual alarm module to issue an audible and visual warning, controlling the instrument panel to display warning information, controlling the onboard airbag ECU to prepare for deployment, and sending a trigger signal to the wearable safety terminal through the encrypted two-way communication link. The vehicle controller simultaneously monitors whether it receives a synchronization trigger signal from the wearable safety terminal. If a synchronization trigger signal is received, it confirms that the wearable safety terminal has also detected a collision risk and continues to execute safety protection actions. Once the collision risk is eliminated, the vehicle controller stops executing safety protection actions and sends a warning cancellation command to the wearable safety terminal.

[0183] The embodiments provided in this application construct a bidirectional redundant triggering mechanism between electric two-wheelers and wearable safety terminals, enabling both electric two-wheelers and wearable safety terminals to have independent collision risk prediction capabilities and safety protection triggering capabilities. If either party detects a collision risk, it can trigger the other party to perform safety protection actions. This solves the problem in related technologies where the triggering of wearable safety terminals relies on single-point judgment, leading to unreliable protection. It also solves the problems of existing wearable safety terminals and vehicle safety systems not being able to link bidirectionally, asynchronous triggering timing, easy failure of single triggering sources, and poor protection effects, thereby improving the reliability and robustness of the safety protection system.

[0184] In one exemplary embodiment, in the prior art, the wearable safety terminal and the vehicle safety system operate independently. When both detect a collision risk simultaneously, they may trigger safety protection actions independently, leading to repeated execution of safety protection actions (such as repeated airbag deployment, repeated triggering of audible and visual warnings, etc.). This not only wastes resources but may also cause unnecessary fright and interference to the rider. Furthermore, when there is a time difference between the trigger signals from the vehicle side and the wearable safety terminal side, it is impossible to determine whether it is the same collision event or different collision events, which may lead to chaotic response from the protection system. To address this deficiency, in this embodiment, by comparing the time difference between the timestamp of receiving the synchronous trigger signal and the timestamp of sending the trigger signal, if the time difference is less than a first preset time interval, it is determined to be the same collision event, and only one safety protection action is executed, avoiding repeated triggering and resource waste.

[0185] In some embodiments, the method further includes: performing at least one security protection action once in response to the time difference between the timestamp of receiving the synchronization trigger signal and the timestamp of sending the trigger signal being less than a first preset time interval.

[0186] In this embodiment, the electric two-wheeler records the transmission timestamp of the trigger signal when sending the trigger signal and the reception timestamp of the synchronous trigger signal when receiving the synchronous trigger signal. The time difference between the transmission timestamp of the trigger signal and the reception timestamp of the synchronous trigger signal represents the difference in detection time between the electric two-wheeler and the wearable safety terminal for the same collision event. If this time difference is less than a first preset time interval, it means that the electric two-wheeler and the wearable safety terminal detected the same collision risk successively within a very short period of time, which is a two-way confirmation of the same collision event, rather than two independent collision events. In this case, the electric two-wheeler performs at least one safety protection action once, that is, the electric two-wheeler does not repeat the triggered safety protection action, but maintains the current execution state of the safety protection action to avoid repeated airbag deployment or repeated triggering of audible and visual warnings. The first preset time interval is a time threshold used to determine whether two trigger events belong to the same collision event. The first preset time interval is determined comprehensively based on factors such as the communication delay between the electric two-wheeler and the wearable safety terminal, the response time difference of the vehicle sensor unit, and the duration of the collision event, and is, for example, set to any value between 100 milliseconds and 500 milliseconds.

[0187] Optionally, when the onboard controller of the electric two-wheeler predicts a collision risk, it generates a trigger signal and records the transmission timestamp T_send of the trigger signal. The onboard controller sends the trigger signal to the wearable safety terminal via an encrypted two-way communication link, and simultaneously executes at least one safety protection action, including at least one of audible and visual warning, instrument display, and airbag deployment. The onboard controller monitors whether it receives a synchronous trigger signal from the wearable safety terminal. If it receives a synchronous trigger signal, it records the reception timestamp T_receive of the synchronous trigger signal. The onboard controller calculates the time difference ΔT between the reception timestamp T_receive and the transmission timestamp T_send = T_receive - T_send. The onboard controller determines whether the time difference ΔT is less than a first preset time interval T_threshold. If not, it determines that the electric two-wheeler and the wearable safety terminal detected different collision events. The onboard controller determines that a new collision event has occurred and updates the safety protection strategy based on the information carried by the synchronous trigger signal, which may require raising the warning level or triggering additional safety protection actions. If so, it is determined that the electric two-wheeler and the wearable safety terminal detected the same collision event. The vehicle controller maintains the current execution state of the safety protection action and does not trigger the safety protection action again, i.e., it executes the safety protection action only once. When the collision risk is eliminated, the vehicle controller stops executing the safety protection action and sends a warning cancellation command to the wearable safety terminal.

[0188] For example, if an electric two-wheeler detects an obstacle ahead while driving, calculates a predicted collision probability of 0.75, which is greater than the preset probability threshold of 0.6, and thus a collision risk is identified, the vehicle's onboard controller generates a trigger signal at time T1 = 1000ms, records the sending timestamp T_send = 1000ms, and sends the trigger signal to a wearable safety terminal, simultaneously executing an audible and visual warning. At time T2 = 1050ms, the wearable safety terminal detects an abnormal rider posture, generates a synchronous trigger signal, and sends it to the electric two-wheeler. At time T3 = 1080ms, the vehicle's onboard controller receives the synchronous trigger signal and records the receiving timestamp T_receive = 1080ms. The onboard controller calculates the time difference ΔT = 1080 - 1000 = 80ms, which is less than the first preset time interval T_threshold = 200ms, determining it as the same collision event, maintaining the current audible and visual warning action, and not triggering it again.

[0189] The embodiments provided in this application compare the time difference between the timestamp of receiving the synchronization trigger signal and the timestamp of sending the trigger signal. When the time difference is less than a first preset time interval, it is determined to be the same collision event, and only one safety protection action is performed. This solves the problem that bidirectional triggering may lead to repeated execution of protection actions in related technologies, and improves the response efficiency and user experience of the safety protection system.

[0190] In an exemplary embodiment, in the prior art, communication between the wearable safety terminal and the vehicle safety system may be delayed or interfered with. When the wearable safety terminal detects a collision risk and sends a synchronization trigger signal, if the signal is delayed during transmission, the electric two-wheeler may only receive the signal after the collision event has ended or the risk has been eliminated. If safety protection actions are still performed at this time, not only will the protection be ineffective, but it may also cause unnecessary fright and interference to the rider. In addition, if the wearable safety terminal generates an incorrect synchronization trigger signal due to a malfunction or misjudgment of the vehicle sensor unit, the electric two-wheeler's blind execution of safety protection actions will lead to false triggering, reducing the user experience. To solve this defect, in this embodiment, by judging whether the time difference between the sending timestamp and receiving timestamp of the synchronization trigger signal is greater than a second preset time interval, and combining it with the current collision risk prediction result of the electric two-wheeler, if the time difference is too large and the electric two-wheeler currently has no collision risk, it is determined that the synchronization trigger signal has expired or is a false trigger, and the safety protection action is not performed, thus avoiding invalid protection responses caused by expired signals and false triggering.

[0191] In some embodiments, the method further includes: in response to the time difference between the timestamp of the wearable safety terminal sending the synchronization trigger signal and the timestamp of the electric two-wheeler receiving the synchronization trigger signal being greater than a second preset time interval, and the predicted collision probability indicating that the electric two-wheeler has no collision risk, then at least one safety protection action is not performed.

[0192] In this system, when the wearable safety terminal generates a synchronization trigger signal, it obtains the current timestamp from the local clock and includes this timestamp as the transmission timestamp in the synchronization trigger signal, sending it to the electric two-wheeler. Upon receiving the synchronization trigger signal, the electric two-wheeler parses the transmission timestamp from the signal, records the reception timestamp, and calculates the difference between the reception and transmission timestamps to obtain the transmission delay time of the synchronization trigger signal.

[0193] If the time difference between the timestamp of the wearable safety terminal sending the synchronization trigger signal and the timestamp of the electric two-wheeler receiving the synchronization trigger signal is greater than the second preset time interval, it means that the synchronization trigger signal has experienced a long delay during transmission. This signal may be for a collision event that has already ended, or its validity may have decreased, rendering it insufficient as a basis for current safety protection decisions. Simultaneously, if the electric two-wheeler predicts that there is no collision risk, it will not execute at least one safety protection action. That is, the electric two-wheeler will not execute a safety protection action only if both of the above conditions are met simultaneously. This dual-judgment mechanism ensures that safety protection actions are only abandoned when the synchronization trigger signal is both expired and inconsistent with the current vehicle state, avoiding misjudgments caused by a single-condition judgment.

[0194] Optionally, the onboard controller of the electric two-wheeler continuously predicts the collision risk of the electric two-wheeler to obtain the current predicted collision probability. The onboard controller monitors whether it receives a synchronization trigger signal sent by the wearable safety terminal. If a synchronization trigger signal is received, the onboard controller parses the transmission timestamp T_send of the synchronization trigger signal sent by the wearable safety terminal from the synchronization trigger signal, and simultaneously records the reception timestamp T_receive of the synchronization trigger signal received by the electric two-wheeler. The onboard controller calculates the time difference ΔT between the reception timestamp T_receive and the transmission timestamp T_send = T_receive - T_send. The onboard controller determines whether the time difference ΔT is greater than a second preset time interval T_threshold2. If not, the onboard controller determines that the synchronization trigger signal is valid and executes at least one safety protection action, including at least one of the following: audible and visual warning, instrument display, and onboard airbag deployment. If so, the vehicle controller determines whether the current predicted collision probability is less than a preset probability threshold, i.e., whether the electric two-wheeler poses a collision risk. If the predicted collision probability is less than the preset probability threshold, i.e., the electric two-wheeler does not pose a collision risk, the vehicle controller determines that the synchronization trigger signal has expired and is inconsistent with the current vehicle status, does not execute any safety protection actions, and discards the synchronization trigger signal. If the predicted collision probability is greater than or equal to the preset probability threshold, i.e., the electric two-wheeler poses a collision risk, at least one safety protection action is executed, including at least one of the following: audible and visual warning, instrument display, and airbag deployment. Once the collision risk is eliminated, the vehicle controller stops executing safety protection actions and sends a warning cancellation command to the wearable safety terminal.

[0195] For example, while an electric two-wheeler is in motion, a wearable safety terminal detects an abnormal rider posture at time T1=1000ms, generates a synchronization trigger signal carrying a transmission timestamp T_send=1000ms, and sends it to the electric two-wheeler. Due to interference in the communication link, the synchronization trigger signal is delayed during transmission, and the electric two-wheeler receives the signal at time T2=2000ms, recording a reception timestamp T_receive=2000ms. The onboard controller calculates the time difference ΔT=2000-1000=1000ms, which is greater than the second preset time interval T_threshold2=500ms. Simultaneously, the onboard controller's current predicted collision probability is 0.3, less than the preset probability threshold of 0.6, meaning the electric two-wheeler currently faces no collision risk. The onboard controller determines that the synchronization trigger signal has expired and is inconsistent with the current vehicle state, does not execute any safety protection actions, and discards the synchronization trigger signal.

[0196] In this embodiment, it is determined whether the time difference between the sending timestamp and the receiving timestamp of the synchronous trigger signal is greater than the second preset time interval. Combined with the current collision risk prediction result of the electric two-wheeler, when the time difference is too large and there is no current collision risk of the electric two-wheeler, no safety protection action is performed. This avoids invalid protection response caused by expired signals and false triggers, and improves the accuracy of the safety protection system and user experience.

[0197] In one exemplary embodiment, the method further includes: in response to predicting that there is no collision risk with the electric two-wheeler and not receiving a synchronization trigger signal from the wearable safety terminal, not performing at least one safety protection action.

[0198] Optionally, the onboard controller of the electric two-wheeler continuously predicts the collision risk of the electric two-wheeler to obtain the current predicted collision probability. The onboard controller determines whether the predicted collision probability is less than a preset probability threshold, i.e., whether the electric two-wheeler has a collision risk. If the predicted collision probability is less than the preset probability threshold, i.e., the electric two-wheeler has no collision risk, the controller monitors whether it receives a synchronization trigger signal sent by the wearable safety terminal. If no synchronization trigger signal is received, it is determined that the electric two-wheeler has no collision risk, and since no synchronization trigger signal has been received from the wearable safety terminal, no safety protection action is performed, and the vehicle status continues to be monitored.

[0199] This embodiment simultaneously determines the collision risk prediction result of the electric two-wheeler and whether a synchronous trigger signal has been received. When both indicate that there is no risk, no safety protection action is performed, thus avoiding false triggering in the absence of risk and improving the accuracy and reliability of the safety protection system.

[0200] In one exemplary embodiment, the method further includes: in response to predicting a collision risk with an electric two-wheeler or receiving a synchronization trigger signal, uploading a warning event record to a cloud server.

[0201] Optionally, the onboard controller of the electric two-wheeler continuously predicts the collision risk of the electric two-wheeler to obtain the current predicted collision probability. The onboard controller determines whether the predicted collision probability is greater than or equal to a preset probability threshold, i.e., whether the electric two-wheeler has a collision risk. If the predicted collision probability is greater than or equal to the preset probability threshold, i.e., the electric two-wheeler has a collision risk, the onboard controller generates a warning event record. The warning event record includes at least one of the following: event timestamp, predicted collision probability, trigger source, vehicle location, vehicle status parameters, and wearable safety terminal status parameters. If the predicted collision probability is less than the preset probability threshold, i.e., the electric two-wheeler does not have a collision risk, the onboard controller monitors whether it receives a synchronization trigger signal sent by the wearable safety terminal. If a synchronization trigger signal is received, the onboard controller generates a warning event record. The warning event record includes at least one of the following: event timestamp, predicted collision probability, trigger source, vehicle location, vehicle status parameters, and wearable safety terminal status parameters. If no synchronization trigger signal is received, the collision risk prediction of the electric two-wheeler continues.

[0202] In this embodiment, whether the collision risk is actively detected by the electric two-wheeler or the synchronous trigger signal is triggered by the wearable safety terminal, the warning event record is uploaded to the cloud server, realizing the cloud archiving and analysis of the warning event, and providing data support for subsequent accident tracing and system optimization.

[0203] In one exemplary embodiment, the method further includes: pushing device status, accident alarms, and rescue progress to a preset channel via a cloud server.

[0204] Among them, "equipment status" refers to the current working status information of the electric two-wheeler and wearable safety terminal, including at least one of the following: online status, battery status, sensor status, communication status, and airbag status. "Accident alarm" refers to the alarm information generated when an electric two-wheeler is involved in a collision or there is a risk of collision, including at least one of the following: accident type, accident level, accident time, accident location, and rider status. "Rescue progress" refers to the execution status of the emergency rescue process after an accident, including at least one of the following: rescue request sending status, rescue response status, rescue personnel location, and estimated arrival time.

[0205] When the electric two-wheeler is powered on or periodically reports its status, the cloud server pushes the device status to the user's app, allowing the user to monitor the real-time operational status of the electric two-wheeler and wearable safety terminal. When the electric two-wheeler detects a collision risk or a collision occurs, the cloud server pushes an accident alert to the user's app, emergency contacts, and rescue platforms, enabling relevant parties to be promptly informed of the accident and take appropriate measures. When the electric two-wheeler triggers an emergency rescue request, the cloud server pushes the rescue progress to the user's app and emergency contacts, allowing relevant parties to track the rescue progress in real time.

[0206] This embodiment pushes device status, accident alarms, and rescue progress to preset channels via a cloud server, realizing the end-to-end transmission of safety protection information. This enables riders, emergency contacts, and rescue platforms to obtain key information in a timely manner, improving the response efficiency of emergency rescue and the user experience.

[0207] In one exemplary embodiment, continuously predicting the collision risk of an electric two-wheeler includes: collecting vehicle sensor parameters of the electric two-wheeler and receiving environmental data of the environment in which the electric two-wheeler is located from a cloud server; predicting the collision risk of the electric two-wheeler based on the vehicle sensor parameters and environmental data to obtain a predicted collision probability; and determining that the electric two-wheeler has a collision risk in response to the predicted collision probability being greater than a preset probability threshold.

[0208] The process of predicting the collision risk of the electric two-wheeler based on vehicle sensor parameters and environmental data to obtain the predicted collision probability has been described in the above embodiments and will not be described again here.

[0209] In one exemplary embodiment, continuously predicting the collision risk of an electric two-wheeler includes: receiving wearable sensor parameters from a wearable safety terminal via near-field communication, determining an abnormal posture score of the target object of the wearable safety terminal based on the wearable sensor parameters, determining a predicted collision probability of the electric two-wheeler colliding based on the abnormal posture score, and determining that the electric two-wheeler has a collision risk in response to the predicted collision probability being greater than a preset probability threshold.

[0210] The wearable sensor parameters of the wearable security terminal have been explained in the above embodiments and will not be repeated here.

[0211] The posture anomaly score of a wearable safety terminal is a comprehensive assessment value used to quantify the degree to which a rider's posture deviates from a normal riding state. The posture anomaly score comprehensively considers the rider's posture angles, acceleration, angular velocity, and other motion parameters. By calculating the deviation of each parameter from a normal riding baseline, a weighted fusion is obtained to obtain a normalized score. The higher the posture anomaly score, the more severe the rider's posture anomaly, and the higher the risk of a collision. The posture anomaly score typically ranges from 0 to 1, where 0 indicates a completely normal posture and 1 indicates a severely abnormal posture.

[0212] In some embodiments, determining the posture anomaly score of the target object of the wearable security terminal based on wearable sensor parameters includes:

[0213] Acquire rider posture angle sequence, acceleration sequence, and angular velocity sequence collected by wearable safety terminal; calculate the variance of posture angle sequence, acceleration sequence, and angular velocity sequence; use the ratio between the posture angle sequence variance and a preset posture angle variance threshold as the posture angle anomaly factor, the ratio between the acceleration sequence variance and a preset acceleration variance threshold as the acceleration anomaly factor, and the ratio between the angular velocity sequence variance and a preset angular velocity variance threshold as the angular velocity anomaly factor; perform a weighted summation of the posture angle anomaly factor, acceleration anomaly factor, and angular velocity anomaly factor to obtain the posture anomaly score of the target object.

[0214] Among them, the variance of the posture angle sequence reflects the degree of fluctuation of the rider's posture angle within a preset time window; the larger the variance, the more unstable the posture. The variance of the acceleration sequence reflects the degree of fluctuation of the rider's acceleration within a preset time window; the larger the variance, the more unstable the motion state. The variance of the angular velocity sequence reflects the degree of fluctuation of the rider's angular velocity within a preset time window; the larger the variance, the more drastic the posture change. The preset posture angle variance threshold, preset acceleration variance threshold, and preset angular velocity variance threshold were obtained by statistically analyzing a large amount of normal riding data, and represent the upper limit of the variance of each parameter under normal riding conditions.

[0215] In other embodiments, determining a posture anomaly score of the target object of the wearable security terminal based on wearable sensor parameters includes:

[0216] The rider's posture angle, acceleration, and angular velocity are acquired by the wearable safety terminal; the posture angle, acceleration, and angular velocity are input into the pre-trained posture anomaly scoring model, and the posture anomaly scoring model outputs a posture anomaly score.

[0217] The posture anomaly scoring model employs a multilayer perceptron model, trained using historical rider posture data. The training data includes multiple sets of rider posture samples, each labeled with a posture anomaly score. The model input consists of parameters such as posture angle, acceleration, and angular velocity, and the output is the posture anomaly score. The posture anomaly scoring model learns the nonlinear mapping relationship between rider posture parameters and the degree of posture anomaly, outputting an accurate posture anomaly score.

[0218] In some embodiments, determining the predicted collision probability of an electric two-wheeler collision based on an attitude anomaly score includes: determining a fourth probability of a collision between the target object of the wearable safety terminal and the target object based on the attitude anomaly score; and fusing the fourth probability with the predicted collision probability obtained based on vehicle sensor parameters and environmental data to obtain the final predicted collision probability.

[0219] In some embodiments, determining a fourth probability of a collision between the wearable security terminal and the target object based on the posture anomaly score includes: calculating the ratio between the posture anomaly score and a preset posture anomaly score threshold to obtain the fourth probability. When the posture anomaly score is less than the preset posture anomaly score threshold, the fourth probability is less than 1; when the posture anomaly score is greater than or equal to the preset posture anomaly score threshold, the fourth probability is equal to 1. The preset posture anomaly score threshold is determined based on statistical analysis of historical data, for example, set to 0.7.

[0220] In some embodiments, determining a fourth probability of a collision between a wearable security terminal and its target object based on an anomaly score includes: inputting the anomaly score into a pre-trained collision probability prediction model, and the collision probability prediction model outputting a fourth probability. The collision probability prediction model employs a logistic regression model or a neural network model, trained using historical anomaly scores and collision event data. The model can learn the mapping relationship between the anomaly score and the collision probability, outputting an accurate fourth probability.

[0221] Optionally, the onboard controller of the electric two-wheeler receives wearable sensor parameters sent by the wearable safety terminal via a near-field communication module. These parameters include motion state data such as the rider's posture angle, acceleration, and angular velocity. The onboard controller extracts the posture angle, acceleration, and angular velocity from the wearable sensor parameters and calculates the deviation of each parameter from the normal riding baseline. Based on the deviations of each parameter, the onboard controller calculates posture angle anomaly factors, acceleration anomaly factors, and angular velocity anomaly factors, and performs a weighted sum of the three to obtain a posture anomaly score for the target object. The onboard controller calculates a fourth probability based on the posture anomaly score; specifically, it calculates the ratio between the posture anomaly score and a preset posture anomaly score threshold. Simultaneously, the onboard controller calculates a vehicle-side predicted collision probability based on vehicle sensor parameters and environmental data. The onboard controller weightedly fuses the fourth probability and the vehicle-side predicted collision probability to obtain the final predicted collision probability. The onboard controller determines whether the final predicted collision probability is greater than a preset probability threshold. If so, it determines that the electric two-wheeler has a collision risk and executes safety protection actions; otherwise, it continues monitoring.

[0222] This embodiment integrates rider posture anomaly scores collected by wearable safety terminals with vehicle sensor parameters to achieve a comprehensive assessment of collision risks from both the rider and vehicle sides. This improves the accuracy and comprehensiveness of collision risk prediction and avoids the problem of missed detections caused by a single data source.

[0223] In one exemplary embodiment, the wearable sensor parameters include various motion state parameters of the target object within a preset time window;

[0224] The preset time window refers to the time interval used to collect motion state parameters. The length of the preset time window is determined comprehensively based on the real-time requirements of collision risk prediction and the computational workload of data processing, and can be set to any value between 1 second and 5 seconds. Within the preset time window, the wearable safety terminal collects motion state parameters at a preset sampling frequency to form time-series data.

[0225] Multiple motion state parameters refer to a set of multi-dimensional parameters collected by wearable safety terminals within a preset time window to characterize the rider's motion state. These parameters include at least two of the following: posture angle, acceleration, angular velocity, rate of change of posture angle, rate of change of acceleration, and rate of change of angular velocity. Each motion state parameter reflects the rider's motion state from different dimensions; combining multiple motion state parameters allows for a more comprehensive assessment of the degree of abnormality in the rider's posture.

[0226] In some embodiments, determining the posture anomaly score of the target object of the wearable safety terminal based on wearable sensor parameters includes: taking each of the multiple motion state parameters as the current motion state parameter, performing the following processing to obtain the anomaly degree corresponding to each motion state parameter: determining the absolute value of the difference between the current motion state parameter and the baseline state parameter under normal riding conditions, and determining the ratio between the absolute value and a preset state parameter threshold as the anomaly degree corresponding to the current motion state parameter; and performing a weighted summation of the anomaly degrees corresponding to each motion state parameter to obtain the posture anomaly score of the target object.

[0227] The current motion state parameters refer to the actual values ​​of a certain type of motion state parameters collected by the wearable safety terminal at the current moment or within the current preset time window. For example, the current attitude angle, current acceleration, and current angular velocity.

[0228] The anomaly score for each motion state parameter is a normalized evaluation value used to quantify the degree to which a single motion state parameter deviates from a normal riding state. The anomaly score is obtained by calculating the ratio of the deviation between the current motion state parameter and the baseline state parameter to a preset state parameter threshold. The value range is usually from 0 to positive infinity, where 0 represents complete normality and a larger value indicates a higher degree of anomaly.

[0229] The baseline state parameters corresponding to each motion state parameter refer to the reference values ​​of various motion state parameters of the rider under normal riding conditions, obtained through statistical analysis of a large amount of normal riding data. Baseline state parameters include baseline posture angle, baseline acceleration, and baseline angular velocity. These baseline state parameters reflect the rider's average motion state under normal riding conditions and serve as a reference standard for assessing postural abnormalities.

[0230] The difference between the current motion state parameters and the baseline parameters under normal riding conditions indicates the degree of deviation of the rider's current motion state from the normal riding state. A positive difference indicates that the current motion state parameters are greater than the baseline value, and a negative difference indicates that the current motion state parameters are less than the baseline value. The larger the absolute value of the difference, the greater the degree of deviation, and the higher the probability of abnormal rider posture.

[0231] Preset state parameter thresholds refer to the maximum allowable deviations of various motion state parameters, used to normalize the degree of deviation into anomaly. These preset state parameter thresholds include attitude angle thresholds, acceleration thresholds, and angular velocity thresholds, determined based on statistical analysis of historical data. They represent the upper limit of the deviation of motion state parameters from the baseline value; exceeding this upper limit indicates a severely abnormal attitude.

[0232] The ratio between the absolute value of the difference between the current motion state parameters and the baseline state parameters under normal cycling conditions and the preset state parameter threshold represents the proportion of the deviation of the current motion state parameters relative to the maximum permissible deviation, i.e., the normalized anomaly. A ratio less than 1 indicates that the deviation is within the permissible range, while a ratio greater than 1 indicates that the deviation exceeds the permissible upper limit. By calculating the ratio, motion state parameters of different dimensions and magnitudes can be unified onto the same evaluation scale, facilitating subsequent weighted fusion.

[0233] Optionally, the onboard controller of the electric two-wheeler receives wearable sensor parameters sent by the wearable safety terminal via a near-field communication module. These wearable sensor parameters include various motion state parameters of the target object within a preset time window. The onboard controller extracts these motion state parameters from the wearable sensor parameters, including attitude angle sequences, acceleration sequences, and angular velocity sequences. The onboard controller uses each of these motion state parameters as the current motion state parameter and performs the following processing: calculating the difference between the current motion state parameter and the baseline state parameter under normal riding conditions, taking the absolute value of the difference, and calculating the ratio between the absolute value and a preset state parameter threshold to obtain the anomaly degree corresponding to the current motion state parameter. The onboard controller performs a weighted summation of the anomaly degrees corresponding to each motion state parameter to obtain the target object's attitude anomaly score. Based on the attitude anomaly score, the onboard controller calculates the predicted collision probability and determines whether the predicted collision probability is greater than a preset probability threshold. If so, it determines that the electric two-wheeler has a collision risk and executes safety protection actions; otherwise, it continues monitoring.

[0234] This embodiment calculates the deviations between various motion state parameters and baseline state parameters, normalizes the deviations to anomaly degree, and then weights and fuses them to achieve a quantitative assessment of the degree of rider posture abnormality. This improves the accuracy and reliability of posture abnormality detection and provides reliable data support for collision risk prediction.

[0235] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided. Optionally, in this embodiment, the above-described safety protection method for an electric two-wheeled vehicle may be applied, but is not limited to, to applications such as... Figure 1 The hardware environment shown includes an electric two-wheeler 102, a cloud server 104, a wearable security terminal 106, and a preset channel 108.

[0236] The electric two-wheeled vehicle safety protection method of this application embodiment is implemented by the electric two-wheeled vehicle 102. Figure 9 This is a schematic flowchart of an optional safety protection method for an electric two-wheeled vehicle according to an embodiment of this application, as shown below. Figure 9 As shown, the process of this method may include the following steps:

[0237] Step S902: In response to the electric two-wheeler being powered on, continuously predict the probability of a collision with the electric two-wheeler to obtain the predicted collision probability;

[0238] Step S904: In response to the predicted collision probability characterizing the electric two-wheeler as having a collision risk, the target probability interval where the predicted collision probability is located is determined from multiple probability intervals, and a trigger signal corresponding to the target risk level corresponding to the target probability interval is generated.

[0239] Step S906: Execute at least one security protection action corresponding to the target risk level and send a trigger signal to the wearable security terminal; the trigger signal is used to instruct the wearable security terminal to execute at least one security response action corresponding to the target risk level.

[0240] The electric two-wheeled vehicle safety protection method in this embodiment can be applied to the field of safety protection for two-wheeled vehicles such as two-wheeled motorcycles and electric two-wheeled vehicles, and can be applied to scenarios where riders monitor collision risks in real time and issue graded warnings during riding.

[0241] Currently, riders primarily rely on independent wearable safety terminals for safety protection. However, the vibration feedback mechanisms used in existing wearable safety terminals are limited by a single mode and a lack of hierarchical differentiation, making it impossible to provide riders with accurate, timely, and actionable risk perception information. This results in low warning effectiveness, high false alarm rates, poor user experience, and difficulty in achieving proactive risk avoidance in complex traffic environments.

[0242] To address the technical problem of single-mode, non-tiered vibration feedback mechanisms in related technologies, which prevent riders from accurately perceiving risks and thus fail to provide adequate hazard avoidance, this embodiment employs a pre-emptive tiered warning mechanism. This mechanism can proactively alert riders to avoid risks, effectively reducing the accident rate. Specifically, by dividing the predicted collision probability into multiple probability intervals, each corresponding to a different risk level, the target risk level is determined based on the probability interval in which the predicted collision probability falls. The corresponding safety protection actions are then executed, and a corresponding trigger signal is sent, achieving tiered warnings that enable riders to take appropriate hazard avoidance measures based on the risk level.

[0243] The process of continuously predicting the probability of collisions involving electric two-wheeled vehicles, and obtaining the predicted collision probability, has been explained in the above embodiments and will not be repeated here.

[0244] In this embodiment, a predicted collision probability greater than a preset probability threshold indicates a collision risk for the electric two-wheeler. After predicting a collision risk, the onboard controller of the electric two-wheeler determines the probability interval in which the predicted collision probability falls from multiple preset probability intervals, designates this probability interval as the target probability interval, obtains the risk level corresponding to the target probability interval as the target risk level, generates a trigger signal corresponding to the target risk level, executes the safety protection action corresponding to the target risk level, and sends the trigger signal to the wearable safety terminal, instructing the wearable safety terminal to execute the safety response action corresponding to the target risk level.

[0245] Here, multiple probability intervals refer to a set of predicted collision probability values ​​divided according to the severity of collision risk. Each probability interval corresponds to a risk level, which refers to the severity of the collision risk represented by that probability interval. The higher the risk level, the more severe the collision risk, and the more urgent the safety protection measures required.

[0246] For example, multiple probability intervals can be divided as follows: the first probability interval is [0.4, 0.6), corresponding to a mild risk level; the second probability interval is [0.6, 0.8), corresponding to a moderate risk level; and the third probability interval is [0.8, 1.0], corresponding to a severe risk level. When the predicted collision probability is 0.55, it falls within the first probability interval, with a target risk level of mild risk; when the predicted collision probability is 0.72, it falls within the second probability interval, with a target risk level of moderate risk; and when the predicted collision probability is 0.88, it falls within the third probability interval, with a target risk level of severe risk.

[0247] In this embodiment, the probability interval in which the predicted collision probability is located among multiple probability intervals is called the target probability interval, and the risk level corresponding to the target probability interval is called the target risk level.

[0248] After predicting a collision risk with an electric two-wheeler, the electric two-wheeler generates a trigger signal corresponding to the target risk level within the target probability range. This signal instructs the wearable safety terminal to execute at least one safety response action corresponding to the target risk level. In other words, the trigger signal carries target risk level information, and the wearable safety terminal executes a safety response action matching that risk level based on this information. Different risk levels correspond to different safety response actions; for example, a minor risk corresponds to a minor vibration warning, a moderate risk to a moderate vibration warning, and a severe risk to the deployment of a wearable airbag. This method achieves a precise match between safety response actions and risk levels, enabling riders to assess the severity of the risk based on the warning intensity and take appropriate avoidance measures.

[0249] Optionally, after the electric two-wheeler is powered on, the onboard controller initializes the onboard sensor unit, communication unit, and execution unit, and begins continuously predicting the probability of a collision. The onboard controller collects vehicle sensor parameters and environmental data, and calculates the predicted collision probability based on these parameters and data. The onboard controller determines whether the predicted collision probability is greater than a preset probability threshold. If not, monitoring continues; if so, it determines that there is a collision risk. The onboard controller then determines the target probability interval from multiple preset probability intervals and identifies the target risk level corresponding to that interval. The onboard controller generates a trigger signal corresponding to the target risk level, carrying at least one of the following: target risk level information, collision risk type information, and warning mode information. The onboard controller executes at least one safety protection action corresponding to the target risk level, including at least one of audible and visual warning, instrument display, and onboard airbag deployment. The onboard controller sends the trigger signal to a wearable safety terminal via an encrypted bidirectional communication link, instructing the wearable safety terminal to execute a safety response action corresponding to the target risk level. Once the collision risk is eliminated, the vehicle controller stops performing safety protection actions and sends a warning cancellation command to the wearable safety terminal.

[0250] The embodiments provided in this application divide the predicted collision probability into multiple probability intervals, each corresponding to a different risk level. The target risk level is determined based on the probability interval in which the predicted collision probability is located, and safety protection actions corresponding to the target risk level are executed and corresponding trigger signals are sent. This achieves graded early warning and solves the technical problem in related technologies where a single-mode, non-graded vibration feedback mechanism leads to riders' inability to accurately perceive risks and failure of risk avoidance. It realizes multi-level, multi-directional, and differentiated vibration feedback based on risk levels, enabling riders to perceive the intensity and direction of risk through touch, significantly improving the accuracy and response efficiency of pre-warning, and enhancing the effectiveness, accuracy, and user experience of early warning.

[0251] In one exemplary embodiment, at least one safety protection action includes at least one of: audible and visual warning, instrument display, and vehicle airbag deployment; multiple probability intervals include a first probability interval, a second probability interval, and a third probability interval; executing at least one safety protection action corresponding to a target risk level includes: in response to the target probability interval being the first probability interval, executing a first safety protection action corresponding to the risk level corresponding to the first probability interval; the first safety protection action includes at least one of the following actions: emitting a warning sound at a first volume, emitting a flashing light of a first color, and displaying a warning icon of a first size on the instrument panel.

[0252] Among them, the sound and light warning is applicable to all risk levels, but the intensity of the sound and light warning is different for different risk levels; the instrument display is applicable to mild and moderate risk levels and is used to provide riders with information on risk type and warning level; the vehicle airbag deployment is only applicable to severe risk levels and is triggered when the collision risk is extremely high and the duration exceeds a preset threshold.

[0253] In this embodiment, multiple probability intervals include a first probability interval, a second probability interval, and a third probability interval. The first probability interval corresponds to a low risk level, the second probability interval corresponds to a moderate risk level, and the third probability interval corresponds to a high risk level. The lower limit of the predicted collision probability for the first probability interval is lower than the lower limit of the predicted collision probability for the second probability interval, and the lower limit of the predicted collision probability for the second probability interval is lower than the lower limit of the predicted collision probability for the third probability interval. A low risk level indicates a low collision risk, giving the rider sufficient reaction time to take evasive action; a moderate risk level indicates a moderate collision risk, requiring the rider to immediately pay attention and take evasive action; a high risk level indicates an extremely high collision risk, requiring the rider to immediately take emergency evasive action or prepare to withstand the impact of a collision.

[0254] When the target probability interval is the first probability interval, the first safety protection action corresponding to the risk level of the first probability interval shall be executed. The first safety protection action refers to a mild warning action taken for a low risk level, which aims to remind the rider of potential risks without causing excessive fright or interference to the rider. The first safety protection action includes at least one of the following actions: issuing a warning sound at a first volume, emitting a flashing light of a first color, or displaying a warning icon of a first size on the instrument panel.

[0255] The first volume level refers to the alert volume corresponding to the mild risk level. The first volume is lower than the second volume corresponding to the moderate risk level and the third volume corresponding to the severe risk level. For example, the first volume is between 50 and 60 decibels. The first volume alert uses a gentle tone, such as a "beep" sound, to remind riders of potential risks.

[0256] The first color refers to the flashing light color corresponding to the mild risk level. The first color uses a less warning color, such as green or blue, to distinguish it from the yellow corresponding to the moderate risk level and the red corresponding to the severe risk level. The first color flashes at a low frequency, such as 1 to 2 times per second, to alert riders to potential risks.

[0257] The first size refers to the size of the warning icon corresponding to a mild risk level displayed on the instrument panel. The first size is smaller than the second size corresponding to a moderate risk level and the third size corresponding to a severe risk level. For example, the first size is 5% to 10% of the instrument panel display area. Warning icons are graphic symbols used to alert riders to the type of risk, such as exclamation marks, warning triangles, and obstacle icons. First-size warning icons are displayed on the instrument panel in a smaller size to remind riders of potential risks without excessively occupying the instrument panel display area.

[0258] Optionally, the vehicle controller determines the target probability interval where the predicted collision probability lies from multiple probability intervals, and checks whether the target probability interval is the first probability interval. If not, it executes the safety protection actions corresponding to other probability intervals; if so, it determines the target risk level as a low risk level and generates a first safety protection action instruction corresponding to the low risk level. The vehicle controller controls the audible and visual alarm module to emit a warning tone at a first volume of 55 decibels, with a "beep" tone, emitted once per second. The vehicle controller controls the lighting module to emit a flashing light of a first color, green, flashing once per second. The vehicle controller controls the instrument display module to display a warning icon of a first size on the instrument panel, the warning icon being an exclamation mark icon, with a display size of 8% of the instrument panel display area. The vehicle controller generates a trigger signal corresponding to the low risk level and sends it to the wearable safety terminal via an encrypted two-way communication link, instructing the wearable safety terminal to execute the first warning mode corresponding to the low risk level.

[0259] This embodiment implements a first safety protection action for mild risk levels, including issuing a warning sound at a first volume, flashing a light of a first color, and displaying a warning icon of a first size on the instrument panel. This achieves a mild warning, reminding riders to pay attention to potential risks, avoiding excessive fright to riders, and improving the humanization of the warning and user experience.

[0260] In an exemplary embodiment, the method further includes: in response to the target probability interval being a second probability interval, performing a second safety protection action corresponding to the risk level corresponding to the second probability interval; the second safety protection action includes at least one of the following actions: emitting a warning sound of a second volume, emitting a flashing light of a second color, and displaying a warning icon of a second size on the instrument panel; the first safety protection action is different from the second safety protection action.

[0261] Specifically, when the target probability interval is the second probability interval, the second safety protection action corresponding to the risk level corresponding to the second probability interval is executed. The second safety protection action refers to a warning action taken for the medium risk level, which aims to force riders to immediately pay attention to the risk and take avoidance measures. The warning intensity is significantly higher than the first safety protection action corresponding to the low risk level. The second safety protection action includes at least one of the following actions: emitting a warning sound at a second volume, emitting a flashing light of a second color, and displaying a warning icon of a second size on the instrument panel.

[0262] The second volume refers to the alert volume corresponding to the moderate risk level. The second volume is greater than the first volume corresponding to the mild risk level. For example, the second volume is between 70 and 80 decibels. The second volume alert uses a rapid tone, such as a "beep beep beep" sound, emitted 2 to 3 times per second, to forcefully remind riders to pay attention to the risk immediately.

[0263] The second color refers to the flashing light color corresponding to the medium-risk level. This second color uses a highly warning color, such as yellow or orange, to distinguish it from the green or blue corresponding to the low-risk level. The second-color light flashes at a higher frequency, such as 3 to 5 times per second, to forcefully remind riders to pay attention to the risk immediately.

[0264] The second size refers to the size of the warning icon corresponding to the moderate risk level displayed on the instrument panel. The second size is larger than the first size corresponding to the mild risk level. For example, the second size is 15% to 25% of the instrument panel display area. The warning icon of the second size is displayed on the instrument panel in a larger size, occupying more of the instrument panel display area, forcibly reminding the rider to pay attention to the risk immediately.

[0265] The first and second safety protection actions differ in at least one of the following ways: The volume of the warning sound is lower than the second; the tone of the warning sound is softer in the first case and more rapid in the second; the light color is green or blue in the first case and yellow or orange in the second; the flashing frequency of the light is lower in the first case and higher in the second; and the size of the warning icon is smaller in the first case than the second. These design differences allow riders to intuitively assess the risk level based on the warning intensity and take appropriate evasive action.

[0266] Optionally, the vehicle controller determines the target probability interval from multiple probability intervals to which the predicted collision probability falls, and checks whether the target probability interval is a second probability interval. If not, it executes safety protection actions corresponding to other probability intervals; if so, it determines the target risk level as a medium risk level and generates a second safety protection action command corresponding to the medium risk level. The vehicle controller controls the audible and visual alarm module to emit a second-volume alert tone, with a volume of 75 decibels and a "beep beep beep" tone, emitted 3 times per second. The vehicle controller controls the lighting module to emit a second-color flashing light, yellow in color, flashing 4 times per second. The vehicle controller controls the instrument display module to display a second-sized warning icon on the instrument panel, a warning triangle icon, with a display size of 20% of the instrument panel display area. The vehicle controller generates a trigger signal corresponding to the medium risk level and sends it to the wearable safety terminal via an encrypted two-way communication link, instructing the wearable safety terminal to execute the second warning mode corresponding to the medium risk level.

[0267] This embodiment implements a second safety protection action for medium-risk levels, including issuing a warning sound at a second volume, flashing lights of a second color, and displaying a warning icon of a second size on the instrument panel. This achieves a warning-type alert, forcibly reminding riders to pay attention to the risk and take avoidance measures, thereby improving the urgency and effectiveness of the warning.

[0268] In an exemplary embodiment, the method further includes: in response to the target probability interval being a third probability interval and the duration of the target probability interval being the third probability interval being greater than a preset duration, executing a third safety protection action corresponding to the risk level corresponding to the third probability interval; the third safety protection action refers to the deployment of the vehicle airbag.

[0269] Specifically, if the target probability interval is the third probability interval, and the duration of the target probability interval being the third probability interval exceeds a preset duration, the third safety protection action corresponding to the risk level of the third probability interval will be executed. The third safety protection action refers to the emergency protection action taken for the severe risk level, namely, the deployment of the vehicle airbag, which aims to provide passive safety protection for the rider before or at the moment of collision, and reduce the injury caused by the collision.

[0270] The duration of the target probability interval being the third probability interval refers to the length of time the predicted collision probability remains within the third probability interval. The duration reflects the duration of the severe collision risk; the longer the duration, the more persistent the collision risk and the higher the probability of a collision. The preset duration is the time threshold used to determine whether to trigger the airbag deployment. The preset duration is determined comprehensively based on factors such as the speed of collision risk development, rider reaction time, and airbag deployment preparation time, and can be set to any value between 0.5 seconds and 2 seconds. When the duration exceeds the preset duration, it indicates that the severe collision risk persists and exceeds the preset threshold, requiring immediate triggering of the airbag deployment to provide passive safety protection for the rider.

[0271] Optionally, the vehicle controller determines the target probability interval from multiple probability intervals to which the predicted collision probability falls. It then checks if the target probability interval is a third probability interval. If not, it executes safety protection actions corresponding to other probability intervals. If so, the vehicle controller records the start time stamp of the predicted collision probability entering the third probability interval and continuously monitors whether the predicted collision probability remains within the third probability interval. The vehicle controller calculates the duration the target probability interval is within the third probability interval and checks if the duration exceeds a preset duration. If so, the vehicle controller determines the target risk level to be a severe risk level and generates an airbag deployment command. The vehicle controller sends the deployment command to the airbag ECU via the vehicle bus, triggering the airbag deployment to provide passive safety protection for the rider. Simultaneously, the vehicle controller executes an audible and visual warning, emitting a loudest warning sound and flashing red lights to alert the rider of an impending collision. The vehicle controller generates a trigger signal corresponding to the severe risk level and sends it to the wearable safety terminal via an encrypted two-way communication link, instructing the wearable safety terminal to deploy the wearable airbag. The vehicle controller uploads collision event records to the cloud server via a wide-area wireless communication module, triggering an emergency rescue request.

[0272] This embodiment achieves emergency protection against severe collision risks by triggering the deployment of the vehicle airbag when the target probability interval is the third probability interval and the duration is longer than the preset duration. It provides passive safety protection for the rider before or at the moment of collision, effectively reducing the damage caused by the collision.

[0273] In one exemplary embodiment, the method further includes: in response to the target probability interval being a third probability interval and the duration of the target probability interval being the third probability interval being less than or equal to a preset duration, performing a second security protection action.

[0274] In this scenario, the target probability interval is defined as the duration of the third probability interval being less than or equal to the preset duration. This means that although the predicted collision probability is within the high-risk range, the duration is short, which could be a momentary risk fluctuation or the rider having already taken effective evasive action to reduce the risk. In this case, the second safety action is executed instead of the third because the duration of the high-risk collision is short and has not yet reached the threshold for triggering airbag deployment. Triggering airbag deployment at this point would be a false trigger, wasting airbag resources and potentially causing unnecessary shock and disturbance to the rider. Executing the second safety action alerts the rider to the risk while avoiding overreaction, giving the rider time and space to take evasive action.

[0275] In this embodiment, when the target probability interval is the third probability interval but the duration is less than or equal to the preset duration, the second safety protection action is executed instead of the third safety protection action. This avoids the false triggering of airbags caused by instantaneous risk fluctuations, and improves the reliability of the system and the user experience while ensuring the effectiveness of the warning.

[0276] In one exemplary embodiment, at least one safety response action includes: performing at least one of body-sensing warning, wearable airbag deployment, and audible and visual warning; multiple probability intervals include a first probability interval and a second probability interval; the risk level corresponding to the first probability interval is lower than the risk level corresponding to the second probability interval.

[0277] In some embodiments, the method further includes: in response to the target probability interval being a first probability interval, determining a trigger signal to instruct the wearable security terminal to execute a first warning mode corresponding to the risk level corresponding to the first probability interval; the first warning mode refers to a pulse vibration mode periodically output with a first vibration intensity, a first vibration frequency, a first single vibration duration, and a first pulse interval duration; in response to the target probability interval being a second probability interval, determining a trigger signal to instruct the wearable security terminal to execute a second warning mode corresponding to the risk level corresponding to the second probability interval; the second warning mode refers to a pulse vibration mode periodically output with a second vibration intensity, a second vibration frequency, a second single vibration duration, and a second pulse interval duration, wherein the first warning mode and the second warning mode are different.

[0278] The tactile warning system is suitable for low and medium risk levels, providing riders with tactile warning signals via a vibration motor. The audible and visual warning system is suitable for low, medium, and high risk levels, providing riders with auditory and visual warning signals via a buzzer and LED lights. The wearable airbag deployment system is only applicable to high risk levels, providing passive safety protection for riders when the collision risk is extremely high. Tactile and audible / visual warnings can be used in combination to form a multimodal warning system, improving the effectiveness of the warnings.

[0279] The first probability interval and the second probability interval have been explained in the above embodiments and will not be repeated here.

[0280] When the target probability interval is the first probability interval, the trigger signal is used to instruct the wearable safety terminal to execute the first warning mode corresponding to the risk level of the first probability interval. After receiving the trigger signal, the wearable safety terminal executes the first warning mode. The first warning mode is a mild pulse vibration mode designed for mild risk levels, which aims to alert riders to potential risks without causing excessive fright or interference to the rider.

[0281] The first vibration intensity refers to the vibration intensity output by the vibration motor in the first warning mode. The first vibration intensity is relatively low, for example, 30% to 50% of the maximum vibration intensity of the vibration motor. The first vibration frequency refers to the output frequency of the pulse vibration in the first warning mode. The first vibration frequency is relatively low, for example, 1 to 2 pulse vibrations per second. The first single vibration duration refers to the duration of each pulse vibration in the first warning mode. The first single vibration duration is relatively short, for example, 100 to 200 milliseconds. The first pulse interval duration refers to the interval between two adjacent pulse vibrations in the first warning mode. The first pulse interval duration is relatively long, for example, 800 to 1000 milliseconds.

[0282] When the target probability interval is the second probability interval, the trigger signal is used to instruct the wearable safety terminal to execute the second warning mode corresponding to the risk level of the second probability interval. After receiving the trigger signal, the wearable safety terminal executes the second warning mode. The second warning mode is a warning pulse vibration mode designed for a medium risk level, which aims to force riders to pay attention to the risk and take avoidance measures. The vibration intensity is significantly higher than that of the first warning mode.

[0283] The second vibration intensity refers to the vibration intensity output by the vibration motor in the second warning mode. The second vibration intensity is higher than the first vibration intensity, for example, 60% to 80% of the maximum vibration intensity of the vibration motor. The second vibration frequency refers to the output frequency of the pulse vibration in the second warning mode. The second vibration frequency is higher than the first vibration frequency, for example, 3 to 5 pulse vibrations per second. The second single vibration duration refers to the duration of each pulse vibration in the second warning mode. The second single vibration duration is longer than the first single vibration duration, for example, 200 to 300 milliseconds. The second pulse interval duration refers to the interval between two adjacent pulse vibrations in the second warning mode. The second pulse interval duration is shorter than the first pulse interval duration, for example, 200 to 400 milliseconds.

[0284] Optionally, the vehicle controller determines the target probability interval where the predicted collision probability lies from multiple probability intervals. If the target probability interval is the first probability interval, the vehicle controller determines the target risk level to be a mild risk level and generates first warning mode parameters, including a first vibration intensity of 40%, a first vibration frequency of 2 times per second, a first single vibration duration of 150 milliseconds, and a first pulse interval duration of 900 milliseconds. The vehicle controller generates a trigger signal carrying the first warning mode parameters and sends it to the wearable safety terminal through an encrypted bidirectional communication link. The wearable safety terminal receives the trigger signal, parses the first warning mode parameters, and controls the vibration motor to periodically output pulse vibrations with a first vibration intensity of 40%, a first vibration frequency of 2 times per second, a first single vibration duration of 150 milliseconds, and a first pulse interval duration of 900 milliseconds.

[0285] If the target probability interval is the second probability interval, the vehicle controller determines the target risk level to be a moderate risk level and generates second warning mode parameters, including a second vibration intensity of 70%, a second vibration frequency of 4 times per second, a second single vibration duration of 250 milliseconds, and a second pulse interval duration of 300 milliseconds. The vehicle controller generates a trigger signal carrying the second warning mode parameters and sends it to the wearable safety terminal through an encrypted two-way communication link. The wearable safety terminal receives the trigger signal, parses the second warning mode parameters, and controls the vibration motor to periodically output pulse vibrations with a second vibration intensity of 70%, a second vibration frequency of 4 times per second, a second single vibration duration of 250 milliseconds, and a second pulse interval duration of 300 milliseconds.

[0286] This embodiment designs different pulse vibration modes for different risk levels, including different vibration intensities, vibration frequencies, duration of a single vibration, and pulse interval durations, achieving graded sensory warnings. This allows riders to intuitively judge the risk level based on vibration intensity and frequency and take corresponding avoidance measures, improving the effectiveness of the warning and the user experience. Understandably, pulse vibrations can also be replaced with other harmless sensory prompts, such as pulsed microcurrents.

[0287] In one exemplary embodiment, the plurality of probability intervals further includes a third probability interval, wherein the risk level corresponding to the second probability interval is lower than the risk level corresponding to the third probability interval; the method further includes: in response to the target probability interval being the third probability interval, determining a trigger signal to instruct the wearable safety terminal to perform wearable airbag deployment.

[0288] Specifically, when the target probability range is the third probability range, the trigger signal instructs the wearable safety terminal to deploy the wearable airbag. Upon receiving the trigger signal, the wearable safety terminal deploys the wearable airbag. The third probability range corresponds to a severe risk level, indicating an extremely high risk of collision, meaning a collision is imminent or is already occurring. In this case, immediate deployment of the wearable airbag is necessary to provide passive safety protection for the rider and mitigate the injuries caused by the collision. The trigger signal carries the airbag deployment command. After receiving the trigger signal, the terminal controller of the wearable safety terminal parses the airbag deployment command, drives the airbag deployment execution module, and triggers the deployment of the wearable airbag, providing cushioning protection for the rider's critical areas (such as the chest, abdomen, and back) at the moment of impact.

[0289] In this embodiment, the wearable airbag is triggered to deploy when the target probability range is the third probability range, thus achieving emergency protection against severe collision risks. It provides passive safety protection for the rider at the moment of collision, effectively reducing the damage caused by the collision and improving the rider's safety.

[0290] In an exemplary embodiment, in the prior art, the vibration warning of wearable safety terminals adopts an indiscriminate somatosensory warning prompt type, which cannot distinguish the direction of the collision source. Riders can only perceive the existence of risk, but cannot determine which direction the risk comes from. This makes it impossible for riders to take targeted avoidance measures (such as avoiding in the opposite direction), resulting in low avoidance efficiency. In fact, riders may even take incorrect avoidance actions due to misjudgment, which may increase the risk of collision.

[0291] To address the aforementioned shortcomings, this embodiment predicts the target collision type of the electric two-wheeler, determines the collision direction based on the collision type, and then determines the target vibration unit on the wearable safety terminal that needs to issue a warning based on the collision direction. This achieves directional warning, enabling riders to determine the direction of the collision source based on the vibration location and take targeted avoidance measures.

[0292] In some embodiments, generating a trigger signal corresponding to a target risk level corresponding to a target probability interval includes: predicting the target collision type of the electric two-wheeler; determining the target collision direction corresponding to the target collision type based on a preset first mapping relationship between collision type and collision direction; determining the target vibration unit on the wearable safety terminal corresponding to the target collision direction based on a preset second mapping relationship between the collision direction and the vibration unit on the wearable safety terminal; generating a trigger signal based on the target risk level and the target vibration unit; the trigger signal is used to instruct the wearable safety terminal to control the target vibration unit to execute a warning mode corresponding to the target risk level.

[0293] The first mapping relationship refers to the correspondence between collision type and collision direction, used to determine the collision direction based on the collision type. The preset collision type refers to the collision category categorized according to the location and manner of the collision, including frontal collision, side collision, rear-end collision, and rollover collision. The collision direction refers to the direction in which the collision force acts or the direction from which the collision originates, including front, rear, left, and right. For example, a frontal collision corresponds to the front, a side collision to the left or right, a rear-end collision to the rear, and a rollover collision to the left or right.

[0294] The target collision type refers to the most likely collision type predicted by the electric two-wheeler based on vehicle sensor parameters and environmental data. The target collision direction refers to the collision direction corresponding to the target collision type, used to indicate which vibration unit on the wearable safety terminal needs to issue a warning.

[0295] In some embodiments, predicting the target collision type for an electric two-wheeler includes:

[0296] Obtain the distance and relative velocity of obstacles in multiple directions around the electric two-wheeler; calculate the predicted collision time between the obstacle and the electric two-wheeler in each direction, where the predicted collision time is the ratio of the absolute value of the obstacle distance to the absolute value of the obstacle relative velocity; determine the direction of the obstacle with the shortest predicted collision time as the target collision direction, and determine the target collision type based on the target collision direction.

[0297] The electric two-wheeler is surrounded by obstacles in multiple directions, including front, rear, left, and right. Obstacle distances are detected using millimeter-wave radar or cameras, and relative speeds are detected using millimeter-wave radar. The predicted collision time reflects the urgency of a collision between the obstacle and the electric two-wheeler; a shorter predicted collision time indicates a more imminent collision. By selecting the direction of the obstacle with the shortest predicted collision time as the target collision direction, the most imminent collision risk source direction can be identified, thus determining the target collision type. For example, if the predicted collision time for the obstacle on the left is 2 seconds and the predicted collision time for the obstacle in front is 5 seconds, then the collision risk for the obstacle on the left is more imminent, the target collision direction is left, and the target collision type is a side collision.

[0298] In other embodiments, predicting the target collision type for the electric two-wheeler includes:

[0299] Obtain the vehicle attitude angle change sequence and vehicle acceleration change sequence of the electric two-wheeler within a preset historical time window; calculate the attitude angle change trend of the vehicle attitude angle change sequence and the acceleration change trend of the vehicle acceleration change sequence; infer the direction of motion instability of the electric two-wheeler based on the attitude angle change trend and acceleration change trend; determine the target collision type based on the direction of motion instability.

[0300] The preset historical time window refers to the time interval used to analyze vehicle motion trends, such as a time window of the most recent 1 to 3 seconds. The vehicle attitude angle change sequence reflects the attitude change process of the electric two-wheeler within the historical time window, and the attitude angle change trend reflects whether the electric two-wheeler has a tendency to tilt to one side. The vehicle acceleration change sequence reflects the acceleration change process of the electric two-wheeler within the historical time window, and the acceleration change trend reflects whether the electric two-wheeler has a tendency to slide to one side.

[0301] In some embodiments, acquiring the vehicle attitude angle change sequence and vehicle acceleration change sequence of an electric two-wheeled vehicle within a preset historical time window includes: acquiring the roll angle, pitch angle, lateral acceleration, and longitudinal acceleration of the electric two-wheeled vehicle at a preset sampling frequency (e.g., 100Hz); performing first-order difference calculation on the roll angle within the preset historical time window to obtain a roll angle change rate sequence, which serves as the vehicle attitude angle change sequence; and performing mean calculation on the lateral acceleration within the preset historical time window to obtain a mean lateral acceleration sequence, which serves as the vehicle acceleration change sequence.

[0302] In some embodiments, calculating the attitude angle change trend of the vehicle attitude angle change sequence includes: performing linear fitting on the roll angle change rate sequence to obtain the slope of the roll angle change rate over time; when the slope is positive and greater than a preset slope threshold, it is determined that the electric two-wheeler is tilting to the right and the attitude angle change trend is tilting to the right; when the slope is negative and less than the negative value of the preset slope threshold, it is determined that the electric two-wheeler is tilting to the left and the attitude angle change trend is tilting to the left; when the absolute value of the slope is less than or equal to the preset slope threshold, it is determined that the electric two-wheeler's attitude is stable and the attitude angle change trend is no tilting trend.

[0303] In some embodiments, calculating the acceleration change trend of the vehicle acceleration change sequence includes: performing sign statistics on the mean lateral acceleration sequence to calculate the proportion of positive values ​​and the proportion of negative values; when the proportion of positive values ​​is greater than a preset proportion threshold, it is determined that the electric two-wheeler is sliding to the right, and the acceleration change trend is sliding to the right; when the proportion of negative values ​​is greater than a preset proportion threshold, it is determined that the electric two-wheeler is sliding to the left, and the acceleration change trend is sliding to the left; when both the proportion of positive values ​​and the proportion of negative values ​​are less than or equal to the preset proportion threshold, it is determined that the lateral motion of the electric two-wheeler is stable, and the acceleration change trend is no sliding trend.

[0304] The direction of motion instability refers to the direction in which an electric two-wheeled vehicle is about to lose balance or overturn, which is inferred by combining the trends of attitude angle change and acceleration change.

[0305] In some embodiments, inferring the direction of motion instability of the electric two-wheeler based on the trends of attitude angle change and acceleration change includes: when the trend of attitude angle change is tilting to the left and the trend of acceleration change is sliding to the left, the direction of motion instability is determined to be to the left; when the trend of attitude angle change is tilting to the right and the trend of acceleration change is sliding to the right, the direction of motion instability is determined to be to the right; when the trends of attitude angle change and acceleration change are inconsistent, the electric two-wheeler is determined to be in an unstable motion state but has not yet lost stability, and the direction of motion instability is uncertain.

[0306] In some embodiments, determining the target collision type based on the direction of motion instability includes: when the direction of motion instability is to the left, the target collision type is a left-side collision or a rollover collision; when the direction of motion instability is to the right, the target collision type is a right-side collision or a rollover collision; when the direction of motion instability is uncertain, the target collision type is determined based on other vehicle sensor parameters (such as the obstacle direction detected by millimeter-wave radar).

[0307] For example, the preset historical time window is 2 seconds, the preset sampling frequency is 100Hz, the preset slope threshold is 5 degrees / second, and the preset proportion threshold is 70%. After linear fitting, the roll angle change rate sequence of the electric two-wheeled vehicle within the most recent 2 seconds has a slope of -8 degrees / second, which is less than the preset slope threshold of -5 degrees / second, indicating a leftward tilt trend in the attitude angle. In the average lateral acceleration sequence of the electric two-wheeled vehicle within the most recent 2 seconds, the negative proportion is 80%, which is greater than the preset proportion threshold of 70%, indicating a leftward sliding trend in the acceleration. Based on this comprehensive assessment, the direction of motion instability is determined to be to the left, and the target collision type is either a left-side collision or a rollover collision.

[0308] The second mapping relationship refers to the correspondence between the collision direction and the vibration units on the wearable safety terminal. This is used to determine which vibration unit needs to issue a warning based on the collision direction. A vibration unit is a device installed on the wearable safety terminal that outputs vibration warnings. Multiple vibration units are located on the front, back, left, and right sides of the wearable safety terminal. The target vibration unit corresponds to the direction of the target collision and is used to indicate the direction of the collision to the rider. For example, the front corresponds to the vibration unit on the front chest, the back to the rear, the left to the left side, and the right to the right side.

[0309] In this embodiment, the trigger signal carries target risk level information and target vibration unit information, which is used to instruct the wearable safety terminal to control the target vibration unit to execute a warning mode corresponding to the target risk level. In this way, directional warning is achieved, enabling riders to determine the direction of the collision source based on the vibration location and take targeted avoidance measures.

[0310] Optionally, the vehicle controller determines the target probability interval from multiple probability intervals to determine the target risk level. The vehicle controller predicts the target collision type based on vehicle sensor parameters and environmental data; for example, based on the obstacle azimuth and relative velocity detected by millimeter-wave radar, it determines the target collision type as a side collision. The vehicle controller determines the target collision direction corresponding to the target collision type based on a first mapping relationship; for example, a side collision corresponds to the left direction. The vehicle controller determines the target vibration unit corresponding to the target collision direction based on a second mapping relationship; for example, the left direction corresponds to the vibration unit on the left side of the wearable safety terminal. The vehicle controller generates a trigger signal based on the target risk level and the target vibration unit. The trigger signal carries target risk level information, target vibration unit identification information, and warning mode parameter information. The vehicle controller sends the trigger signal to the wearable safety terminal via an encrypted bidirectional communication link. The wearable safety terminal receives the trigger signal, parses the target vibration unit identification and warning mode parameters, and controls the vibration unit on the left side to execute the warning mode corresponding to the target risk level.

[0311] This embodiment predicts the target collision type and determines the target collision direction, thereby identifying the target vibration unit and realizing directional early warning. This enables riders to judge the direction of the collision source based on the vibration location and take targeted avoidance measures, improving the accuracy and practicality of the early warning.

[0312] In one exemplary embodiment, the method further includes: in response to the collision probability indicating the elimination of the collision risk of the electric two-wheeler, stopping the execution of at least one safety protection action and sending a warning cancellation command to the wearable safety terminal; the warning cancellation command is used to instruct the wearable safety terminal to stop executing at least one safety response action.

[0313] Optionally, the onboard controller determines whether the predicted collision probability is less than a preset risk clearance threshold. This preset risk clearance threshold, being less than a preset probability threshold, is used to determine whether the collision risk has been completely eliminated. For example, the preset probability threshold is 0.6, and the preset risk clearance threshold is 0.3. If the predicted collision probability is less than the preset risk clearance threshold, the collision risk is determined to be cleared, and current safety protection actions are stopped, including stopping audible and visual warnings, stopping the display of warning icons on the instrument panel, and stopping airbag deployment preparation. The onboard controller generates a warning clearance command, which carries at least one of the following: a risk clearance timestamp, a reason for risk clearance, and current vehicle status information. The onboard controller sends the warning clearance command to the wearable safety terminal via an encrypted two-way communication link. The wearable safety terminal receives the warning clearance command and stops executing current safety response actions, including stopping motion-sensor warnings, stopping audible and visual warnings, and canceling airbag deployment preparation. The wearable safety terminal sends a warning clearance confirmation signal to the electric two-wheeler, confirming that the safety response actions have been stopped. The electric two-wheeler's onboard controller records the risk clearance event and resumes normal monitoring.

[0314] In this embodiment, after the collision risk is eliminated, the safety protection action is stopped and a warning cancellation command is sent to the wearable safety terminal, which realizes the timely cancellation of the warning, avoids the interference caused to the rider by the continuous warning after the risk is eliminated, and improves the intelligence level of the system and the user experience.

[0315] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided. Optionally, in this embodiment, the above-described safety protection method for an electric two-wheeled vehicle may be applied, but is not limited to, to applications such as... Figure 1 The hardware environment shown includes an electric two-wheeler 102, a cloud server 104, a wearable security terminal 106, and a preset channel 108.

[0316] The electric two-wheeled vehicle safety protection method of this application embodiment is implemented by the electric two-wheeled vehicle 102. Figure 10 This is a schematic flowchart of an optional safety protection method for an electric two-wheeled vehicle according to an embodiment of this application, as shown below. Figure 10 As shown, the process of this method may include the following steps:

[0317] Step S1002: In response to the electric two-wheeler being powered on, scan the wearable security terminal in the environment where the electric two-wheeler is located, obtain the identity identifier of the wearable security terminal, and obtain the identity and authentication key of the wearable security terminal that is bound to the identification information of the electric two-wheeler.

[0318] Step S1004: In response to the matching of the identity of the wearable security terminal with the identity of the pairable wearable security terminal, an authentication challenge message encrypted with the authentication key is sent to the wearable security terminal; the authentication challenge message is used to instruct the wearable security terminal to return an authentication response message;

[0319] Step S1006: Receive the authentication response message returned by the wearable security terminal, decrypt the authentication response message using the authentication key, and verify the authentication response message;

[0320] Step S1008: In response to the successful authentication response message, an encrypted communication link is established between the electric two-wheeler and the wearable security terminal.

[0321] The electric two-wheeled vehicle safety protection method in this embodiment can be applied to the field of safety protection for two-wheeled vehicles such as two-wheeled motorcycles and electric two-wheeled vehicles, and can be applied to the scenario of identity authentication and communication link establishment between the rider's wearable safety terminal and the electric two-wheeled vehicle before riding.

[0322] Currently, riders primarily use standalone wearable safety terminals for safety protection. However, existing linkage solutions mostly employ one-way communication, with latency generally exceeding 200ms, failing to meet the stringent timeliness requirements of collision protection. Furthermore, the lack of a two-way device authentication mechanism makes them prone to false triggering, unauthorized access, and signal tampering, resulting in insufficient stability during riding. In addition, the pairing and binding process between wearable safety terminals and electric two-wheelers in existing technologies is cumbersome, requiring manual operation by the user, and lacks an effective authentication mechanism, failing to ensure that only the bound wearable safety terminal can establish communication with the electric two-wheeler, posing a safety hazard.

[0323] In summary, the communication between electric two-wheelers and wearable safety terminals has the following problems: First, there is a lack of identity authentication mechanism, allowing any wearable safety terminal to establish communication with the electric two-wheeler, posing a risk of unauthorized device access; second, there is a lack of encryption protection, making communication content easily eavesdropped on and tampered with, posing a risk of data leakage; third, the pairing process is cumbersome, requiring manual operation by the user, resulting in a poor user experience; and fourth, the authentication process relies on the network, making authentication impossible in environments without a network, thus limiting system availability.

[0324] To address the challenges of secure authentication and communication link establishment between wearable security terminals and electric two-wheelers, and to achieve secure, convenient, and automated identity authentication and communication link establishment between them, this embodiment pre-establishes a three-way binding relationship between the user account, the wearable security terminal, and the electric two-wheeler via a cloud server. A unique authentication key is then generated and distributed to the electric two-wheeler. Upon powering on, the electric two-wheeler automatically scans for nearby wearable security terminals and performs two-way identity authentication using the authentication key. Once authentication is successful, an encrypted communication link is automatically established, thus achieving secure, convenient, and automated identity authentication and communication link establishment.

[0325] In this embodiment, a three-element binding and two-way authentication mechanism is established between the user, the wearable safety terminal, and the electric two-wheeler to ensure secure device access and prevent unauthorized access and accidental triggering. It is also compatible with two-wheelers of different brands and protocols, supporting OTA remote upgrades and resolving the issues of poor adaptability and lack of iteration in existing solutions. Furthermore, universal wireless communication technology enables two-way encrypted communication, covering all mainstream wireless connection technologies to ensure high reliability and low latency data transmission. The three-element binding and two-way authentication mechanism effectively prevents unauthorized device access and malicious tampering of trigger signals. A local redundant storage mechanism ensures the integrity and traceability of accident data, providing core evidence for accident liability determination.

[0326] The three-way binding process for user account, wearable safety terminal, and electric two-wheeler includes: the user scans the identity identifier of the wearable safety terminal (such as the serial number QR code of the vest terminal) through a mobile APP to obtain the terminal's unique identity information (such as the vest's unique serial number and Bluetooth MAC address); the user enters the vehicle's unique identifier (such as the VIN code) in the APP and submits it to the cloud management platform to complete the three-way binding of user account, wearable safety terminal, and vehicle, and the cloud generates a unique binding authentication key; the cloud platform sends the bound terminal identity information and authentication key to the corresponding vehicle's onboard safety system (onboard T-BOX); the onboard system initiates near-field communication scanning, matches the wearable safety terminal with the corresponding identity, completes two-way identity authentication through the key, and establishes an encrypted two-way communication link; each time the vehicle is powered on, the onboard system and wearable safety terminal automatically complete two-way authentication and link reconnection, without requiring manual operation by the user.

[0327] The cloud server stores three-element binding information, which refers to the binding relationship between user account information, wearable security terminal identity identifier, and electric two-wheeler identifier information. This information is used to ensure that only bound users can obtain full system function permissions when using the bound wearable security terminal to operate the bound electric two-wheeler.

[0328] After the electric two-wheeler is powered on, its onboard controller activates the near-field communication module to scan for nearby wearable safety terminals and obtain their identification information. The wearable safety terminal's identification information refers to the unique identification information used to identify the wearable safety terminal, including at least one of the following: device serial number (SN code), Bluetooth MAC address, and unique device identifier. The electric two-wheeler's identification information refers to the unique identification information used to identify the electric two-wheeler, including at least one of the following: vehicle identification number (VIN code), vehicle serial number, and unique vehicle identifier.

[0329] The identity of a pairable wearable safety terminal refers to the identity identifier of the wearable safety terminal that is bound to the current electric two-wheeler's identification information. In other words, it is the identity identifier of the wearable safety terminal that is allowed to pair and communicate with the electric two-wheeler, as determined by the cloud server based on the three-element binding information.

[0330] An authentication key is an encryption key generated by a cloud server based on ternary binding information, used for two-way authentication between the electric two-wheeler and the wearable security terminal. The authentication key ensures the authenticity of the identities of both parties and the confidentiality of the communication content, preventing unauthorized device access and signal tampering. For example, if the authentication key is a symmetric encryption key, the cloud server uses the Advanced Encryption Standard (AES) algorithm to generate a 256-bit symmetric key and distributes this key to both the electric two-wheeler and the wearable security terminal. Both the electric two-wheeler and the wearable security terminal use the same key for encryption and decryption. Alternatively, if the authentication key is an asymmetric encryption key pair, the cloud server generates a public key and a private key pair, distributes the public key to the electric two-wheeler, and the private key to the wearable security terminal. The electric two-wheeler uses the public key to encrypt the authentication challenge message, and the wearable security terminal uses the private key to decrypt the authentication challenge message and generate an authentication response message.

[0331] In some embodiments, the methods for obtaining the identity and authentication key of the pairable wearable security terminal bound to the identification information of the electric two-wheeler include the following three methods:

[0332] Method 1: Obtain locally from the vehicle. The electric two-wheeler's onboard controller reads the pre-stored identity and authentication key of the pairable wearable security terminal from its local secure storage area. This pairable wearable security terminal identity and authentication key is then distributed to the electric two-wheeler by the cloud server after the three-factor binding is completed. The electric two-wheeler stores it in its local secure storage area (such as a security chip or encrypted storage area). The advantage of local storage on the vehicle is that authentication can be completed without a network connection, resulting in fast response times, making it suitable for authentication scenarios in environments without a network connection.

[0333] Method 2: Obtaining information locally from the wearable security terminal. The electric two-wheeler sends an identity query request to the wearable security terminal via its near-field communication module. The wearable security terminal reads the pre-stored identity and authentication key of the pairable electric two-wheeler from its local secure storage area and returns the pairable electric two-wheeler's identity to the electric two-wheeler. The electric two-wheeler verifies whether the returned pairable electric two-wheeler's identity matches its own identification information. If they match, it obtains the authentication key for subsequent authentication. The advantage of local storage on the wearable device is that the wearable security terminal can proactively provide identity information, facilitating the electric two-wheeler's quick identification of the bound device.

[0334] Method 3: Obtaining from a cloud server. The electric two-wheeler sends an identity query request to the cloud server via its wireless communication module. The cloud server queries the three-element binding information based on the electric two-wheeler's identification information and returns the identity and authentication key of the pairable wearable security terminal. The advantage of obtaining information from the cloud is that the binding information can be updated in real time, supporting dynamic replacement of wearable security terminals. However, the most obvious drawback is that in an environment without a network, the electric two-wheeler cannot connect to the cloud server, resulting in the inability to obtain the identity and authentication key of the pairable wearable security terminal, and the authentication process cannot proceed. Therefore, local storage on both the vehicle and the wearable is a more reliable implementation method, ensuring that identity authentication and communication link establishment can be completed even in an environment without a network.

[0335] In this process, the wearable security terminal's identity matches the identity of the pairable wearable security terminals. This means that the scanned wearable security terminal is a device pre-bound by the cloud server and allowed to pair and communicate with the electric two-wheeler, rather than an illegal or unbound device. In this case, the electric two-wheeler sends an authentication challenge message encrypted with the authentication key to the wearable security terminal. The authentication challenge message is an encrypted message generated by the electric two-wheeler to verify the authenticity of the wearable security terminal's identity, instructing the wearable security terminal to return an authentication response message to complete two-way authentication. After receiving the authentication challenge message, the wearable security terminal decrypts the authentication challenge message using its locally stored authentication key, generates an authentication response message based on the decrypted content, and returns the authentication response message to the electric two-wheeler. The authentication response message is an encrypted message generated by the wearable security terminal to prove its own authenticity, responding to the electric two-wheeler's authentication challenge and completing two-way authentication.

[0336] After receiving the authentication response message from the wearable security terminal, the electric two-wheeler decrypts and verifies the message to ensure that it indeed comes from the bound wearable security terminal and that the message content has not been tampered with, thus completing two-way identity authentication.

[0337] In some embodiments, decrypting the authentication response message using the authentication key and verifying the authentication response message includes: the on-board controller of the electric two-wheeler parses the authentication response message to obtain the timestamp, random number, encrypted terminal identifier, and challenge hash value of the authentication response message sent by the wearable security terminal; the on-board controller verifies whether the timestamp is within a preset valid time window; if the timestamp exceeds the valid time window, it is determined that the authentication response message has expired and the verification fails; the on-board controller uses the authentication key to derive a temporary decryption key, and uses the temporary decryption key to decrypt the encrypted terminal identifier to obtain the identity identifier of the wearable security terminal; the on-board controller verifies whether the identity identifier of the decrypted wearable security terminal is consistent with the locally stored identity of the pairable wearable security terminal; if they are consistent, the verification passes; if they are inconsistent, the verification fails.

[0338] In some embodiments, decrypting the authentication response message using an authentication key and verifying the authentication response message includes: the on-board controller of the electric two-wheeler parses the authentication response message to obtain the random number return value and signature value returned by the wearable security terminal; the on-board controller verifies whether the random number return value is consistent with the random number carried in the authentication challenge message; if they are inconsistent, it is determined that the authentication response message has been tampered with or comes from an unauthorized device, and the verification fails; the on-board controller uses the authentication key to verify the correctness of the signature value; if the signature verification passes, it is confirmed that the authentication response message comes from the bound wearable security terminal; if the signature verification fails, it is determined that the authentication response message has been tampered with, and the verification fails; when both the random number return value and the signature verification pass, it is determined that the authentication response message has been verified successfully.

[0339] The successful authentication response message confirms the authenticity of the wearable security terminal, demonstrating that it is indeed a legitimate device pre-bound to the cloud server and that the messages were not tampered with during communication. In this case, an encrypted communication link is established between the electric two-wheeler and the wearable security terminal.

[0340] In some embodiments, in response to an authentication response message failing to verify, an encrypted communication link between the electric two-wheeler and the wearable security terminal is refused to be established.

[0341] Optionally, after the electric two-wheeler is powered on, the onboard controller initializes the near-field communication module and starts scanning for nearby wearable safety terminals. The onboard controller acquires the identity identifiers of the scanned wearable safety terminals, including the device serial number and Bluetooth MAC address. The onboard controller reads the identity and authentication key of the pairable wearable safety terminal bound to the electric two-wheeler's identification information from its local secure storage area. This authentication key is issued and stored by the cloud server. The onboard controller compares the identity identifier of the scanned wearable safety terminal with the identity of the pairable wearable safety terminal to determine if they match. If the identity identifier matches the identity of the pairable wearable safety terminal, the onboard controller generates an authentication challenge message. The onboard controller encrypts the authentication challenge message using the authentication key and sends it to the wearable safety terminal through the near-field communication module. The onboard controller waits to receive the authentication response message from the wearable safety terminal. If no authentication response message is received within a preset timeout period, authentication is considered a failure. After receiving the authentication response message, the onboard controller decrypts and verifies the authentication response message using the authentication key. If the authentication response message is successfully verified, the vehicle controller establishes an encrypted communication link with the wearable security terminal and begins normal data interaction. If the authentication response message fails to verify, the vehicle controller refuses to establish a communication link and prompts the user of authentication failure via the instrument panel or voice.

[0342] The embodiments provided in this application establish a three-way binding relationship between a user account, a wearable security terminal, and an electric two-wheeler on a cloud server, generating a unique authentication key and distributing it to both the electric two-wheeler and the wearable security terminal. This three-way binding relationship ensures that only the bound user, using the bound wearable security terminal to operate the bound electric two-wheeler, can obtain full system access, thus eliminating the possibility of unauthorized device access at the source. Secondly, after powering on, the electric two-wheeler automatically scans for nearby wearable security terminals, obtains their identity identifiers, and matches them with locally stored pairs of wearable security terminal identities. Only wearable security terminals with matching identities can proceed to the subsequent authentication process, achieving initial device identity screening. Then, the electric two-wheeler sends an authentication challenge message encrypted with the authentication key to the wearable security terminal. The wearable security terminal uses the locally stored authentication key to decrypt the authentication challenge message and generates an authentication response message, which is returned to the electric two-wheeler. The electric two-wheeler decrypts and verifies the authentication response message; upon successful verification, an encrypted communication link is established. Through two-way authentication and encrypted communication, the authenticity of both parties' identities and the confidentiality of the communication content are ensured. Finally, the authentication key and the identity of the pairable wearable security terminal can be stored locally on the electric two-wheeler or the wearable security terminal, enabling authentication without a network connection and ensuring system availability in offline environments. In summary, the above authentication process resolves the technical problem of the risk of unauthorized device access between the electric two-wheeler and the wearable security terminal in related technologies.

[0343] In one exemplary embodiment, the authentication key is generated by the cloud server based on the ternary binding information between the object account information, the identity of the wearable security terminal, and the identification information of the electric two-wheeler.

[0344] The object account information refers to the account information registered by the user on the cloud server, including at least one of the following: user unique identifier, user name, user mobile phone number, and user ID card number. The object account information is used to uniquely identify the user using the wearable safety terminal and the electric two-wheeler.

[0345] The three-element binding process involves a user scanning the identification identifier of a wearable security terminal (such as the serial number QR code on a vest-like device) via a mobile app to obtain the terminal's unique identification information (such as the vest's unique serial number and Bluetooth MAC address). The user then enters the vehicle's unique identifier (such as the VIN code) within the app and submits it to the cloud management platform, completing the three-element binding between the user account, wearable security terminal, and vehicle. A unique binding authentication key is generated in the cloud. The cloud platform then distributes the bound terminal identification information and authentication key to the corresponding vehicle's onboard safety system (vehicle T-BOX). The three-element binding information includes the binding correspondence between user account information, wearable security terminal identification identifier, and electric two-wheeler identification information, ensuring that only the bound user can obtain full system function permissions when using the bound wearable security terminal to operate the bound electric two-wheeler.

[0346] In some embodiments, the cloud server generates an authentication key based on a three-factor binding information between the object account information, the identity of the wearable security terminal, and the identification information of the electric two-wheeler, including:

[0347] The cloud server obtains the user's unique identifier from the user's account information, the device serial number from the wearable security terminal's identity identifier, and the vehicle identification number from the electric two-wheeler's identification information; the cloud server concatenates the user's unique identifier, device serial number, and vehicle identification number in a preset order to obtain a concatenated string; the cloud server performs a hash operation on the concatenated string using a preset hash algorithm to obtain a hash value; the cloud server uses the first 256 bits of the hash value as the authentication key.

[0348] The preset order can be user unique identifier first, device serial number second, vehicle identification number last, or any other preset order. The preset hash algorithm can be a secure hash algorithm (SHA-256) or a message digest algorithm (MD5), etc. By concatenating the three-element binding information and then performing a hash operation, it is ensured that the authentication key uniquely corresponds to the three-element binding information, and the three-element binding information cannot be deduced from the authentication key, thus improving the security of the authentication key.

[0349] In other embodiments, the cloud server generates an authentication key based on a three-factor binding information between the object's account information, the wearable security terminal's identity identifier, and the electric two-wheeler's identification information, including:

[0350] The cloud server obtains the user's mobile phone number from the user's account information, the Bluetooth MAC address from the wearable security terminal's identification information, and the vehicle serial number from the electric two-wheeler's identification information; the cloud server generates a random number as a key seed; the cloud server takes the user's mobile phone number, Bluetooth MAC address, vehicle serial number, and key seed as input parameters, inputs them into a preset key derivation function, and the key derivation function outputs the authentication key.

[0351] The key derivation function can be a hash-based message authentication code key derivation function (HKDF) or a cryptographic key derivation function (PBKDF2), etc. The key derivation function can convert multiple input parameters into a fixed-length key, and the output is cryptographically secure. By introducing a random number as the key seed, the randomness of each generated authentication key is ensured. Even if the three-element binding information is the same, the authentication keys generated at different times will be different, improving the authentication key's resistance to replay attacks.

[0352] In this embodiment, the cloud server generates an authentication key based on the three-element binding information, ensuring that the authentication key uniquely corresponds to the binding relationship between the user account, the wearable security terminal, and the electric two-wheeler. Only the bound device can pass the identity authentication, thereby improving the system's security and anti-counterfeiting capabilities.

[0353] In one exemplary embodiment, the method further includes: in response to receiving a new wearable security terminal identity and a new authentication key issued by a cloud server, deleting the locally stored wearable security terminal identity and authentication key, and storing the new wearable security terminal identity and the new authentication key, wherein the new wearable security terminal identity and the new authentication key are generated by the cloud server in response to a wearable security terminal replacement request from a mobile application terminal.

[0354] In practical applications, there are various scenarios where the three-element binding information needs to be updated. For example, a user's original wearable security terminal may be damaged, lost, or upgraded, requiring a new one; a user may lend their wearable security terminal to another user, requiring rebinding; or a user may purchase a new electric two-wheeler, requiring the original wearable security terminal to be bound to the new vehicle. In these scenarios, users need to initiate a wearable security terminal replacement request through a mobile application.

[0355] A wearable security terminal replacement request refers to a user's request sent from a mobile application terminal to the cloud server to change the currently bound wearable security terminal. This request updates the three-way binding relationship between the user account, the wearable security terminal, and the electric two-wheeler. The purpose of the wearable security terminal replacement request is to trigger the cloud server to generate new three-way binding information and a new authentication key, ensuring that the new wearable security terminal can establish secure communication with the electric two-wheeler.

[0356] In this scenario, after receiving a wearable security terminal replacement request from a mobile application terminal, the cloud server parses the request to obtain the identity of the new wearable security terminal, user account information, and electric two-wheeler identification information. Based on the new wearable security terminal's identity, user account information, and electric two-wheeler identification information, the cloud server generates a new three-element binding relationship. Based on the new three-element binding relationship, the cloud server generates a new authentication key. Finally, the cloud server sends the new wearable security terminal's identity and the new authentication key to the electric two-wheeler's onboard system.

[0357] After receiving the new wearable security terminal identity and new authentication key from the cloud server, the on-board system of the electric two-wheeler deletes the original wearable security terminal identity and original authentication key stored locally, stores the new wearable security terminal identity and new authentication key, and completes the update of the three-element binding information.

[0358] The new wearable safety terminal identity refers to the identity identifier of the new wearable safety terminal after the user replaces it, including at least one of the following: device serial number, Bluetooth MAC address, and unique device identifier. The new authentication key refers to the encryption key generated by the cloud server based on the updated three-element binding relationship, used for two-way identity authentication between the electric two-wheeler and the new wearable safety terminal.

[0359] Optionally, the onboard controller of the electric two-wheeler receives the new wearable security terminal identity and new authentication key from the cloud server via a wireless communication module. The onboard controller verifies the integrity and validity of the new wearable security terminal identity and new authentication key, including verifying the correct data format, valid signature, and valid timestamp. The onboard controller accesses the local secure storage area and reads the currently stored wearable security terminal identity and authentication key. The onboard controller deletes the currently stored wearable security terminal identity and authentication key from the local secure storage area. The onboard controller writes the new wearable security terminal identity and new authentication key to the local secure storage area. The onboard controller updates the local binding status flag, marking that the three-element binding information has been updated. The onboard controller prompts the user via the instrument panel or voice to confirm the successful replacement of the wearable security terminal. Upon the next power-on, the onboard controller uses the new wearable security terminal identity and new authentication key to authenticate with the new wearable security terminal and establish a communication link.

[0360] In this embodiment, the cloud server generates new three-element binding information and authentication key in response to the replacement request of the mobile application terminal. The electric two-wheeler automatically updates the locally stored identity information and authentication key, realizing convenient replacement of wearable security terminal and improving system flexibility and user experience.

[0361] In one exemplary embodiment, the above method includes: recording security logs of all authentication operations and permission change operations, the security logs including operation time, operation type, operation user, and operation result; uploading the security logs to a cloud server; recording the security logs through the cloud server; and querying and monitoring the account security status of specified object account information through a preset channel.

[0362] The authentication operation refers to the identity authentication-related operations between the electric two-wheeler and the wearable security terminal, including at least one of the following: sending an authentication challenge message, receiving an authentication response message, verifying the authentication response message, authentication success, authentication failure, establishing an encrypted communication link, and disconnecting the encrypted communication link. The authentication operation is used to record key events in the identity authentication process for easy retrospective tracking and security auditing.

[0363] Permission change operations refer to operations related to the three-way binding relationship between user accounts, wearable security terminals, and electric two-wheelers. These include at least one of the following operations: replacement of wearable security terminals, updating authentication keys, binding user accounts, unbinding user accounts, granting device permissions, and revoking device permissions. Permission change operations are used to record key events in the permission management process, ensuring the traceability of permission changes.

[0364] Security logs are records of detailed information regarding authentication and permission change operations. They store critical data from the operational process for easy retrieval, auditing, and analysis later. Security logs are stored encrypted to ensure the integrity and immutability of their content.

[0365] The security log records the operation time, operation type, operating user, and operation result. Operation time refers to the specific point in time when the authentication or permission change operation occurred. Operation time identifies the temporal sequence of operations, facilitating the analysis of their time series and correlations. Operation type refers to the specific type of authentication or permission change operation, such as "authentication challenge sending," "authentication response verification," "wearable security terminal replacement," and "authentication key update." Operation type is used to categorize and manage different types of operation records, facilitating querying and statistics by type. Operating user refers to the user's identity identifier who initiated the authentication or permission change operation, including at least one of the user's unique identifier, user name, or user mobile phone number. The operating user identifies the initiator of the operation, facilitating the tracing of the responsible party. Operation result refers to the execution result of the authentication or permission change operation, including at least one of success, failure, timeout, and rejection. The operation result identifies the final state of the operation, facilitating the statistical analysis of success rate and reasons for failure. Specified target account information refers to the target user account information that the user needs to query and monitor, including at least one of the user's unique identifier, user name, and user mobile phone number. Specifying target account information is used to define the target scope for querying and monitoring, allowing users or administrators to view the security status of specific accounts. Account security status refers to the security assessment result of a user account, including at least one of the following: normal, abnormal, risk warning, and security alert. Account security status is comprehensively assessed based on authentication and permission change records in the security logs, used to alert users or administrators to potential security risks.

[0366] Optionally, the onboard controller of the electric two-wheeler monitors the occurrence of authentication and permission change operations. When an authentication or permission change operation is detected, the onboard controller obtains the operation time, operation type, user, and result. The onboard controller combines the operation time, operation type, user, and result to generate a security log entry. The onboard controller writes the security log entry to the local security log storage area and stores it using encryption. The onboard controller determines whether the security log upload conditions are met, including at least one of the following: normal network connection, insufficient local storage space, or reaching a preset upload period. If the upload conditions are met, the onboard controller uploads the locally stored security log to the cloud server via the wireless communication module. The cloud server receives the security log, stores it in the cloud database, and establishes an index associated with user account information. Users or administrators log in to the cloud server through a mobile application terminal or web management platform and enter the specified object account information to query. The cloud server queries the associated security logs based on the specified object account information and analyzes the authentication and permission change operation records. Based on the security log analysis results, the cloud server assesses the account security status of the specified object account information and generates a security status report. The cloud server returns the account security status and security status report to the mobile application or web management platform for display. If the account security status is abnormal or a risk warning is issued, the cloud server sends a security alert notification to the user or administrator.

[0367] This embodiment records security logs of authentication and permission change operations and uploads them to the cloud server, enabling full traceability and security auditing of operational behavior. Users and administrators can query and monitor account security status in real time, promptly identify and respond to security risks, and improve the security and manageability of the system.

[0368] In an exemplary embodiment, the authentication process between an electric two-wheeler and a wearable security terminal in the prior art has the following drawbacks: First, authentication messages are easily intercepted and replayed, allowing attackers to intercept legitimate authentication messages and repeatedly send them, leading to unauthorized device access; second, the content of authentication messages is easily tampered with, allowing attackers to modify key information and impersonate the user; third, the keys used in the authentication process remain unchanged for a long time, posing a risk of key leakage, which compromises the security of the entire system. To address these drawbacks, this embodiment introduces techniques such as timestamps, random numbers, key derivation functions, temporary keys, and challenge hash values ​​to achieve anti-replay, anti-tampering, and dynamic key updates for authentication messages, thereby improving the security of the authentication process.

[0369] In some embodiments, sending an authentication challenge message encrypted with an authentication key to a wearable security terminal includes: obtaining a current timestamp as a first timestamp, generating a first random number and a vehicle-side random number, and obtaining locally stored identification information of an electric two-wheeler; based on a preset key derivation function, using the authentication key as a master key, deriving a first temporary encryption key and a first temporary signature key according to the first random number; encrypting the identification information of the electric two-wheeler using the first temporary encryption key and the vehicle-side random number to obtain an encrypted vehicle identifier; generating a first challenge hash value according to the first temporary signature key, the first timestamp, the first random number, and the encrypted vehicle identifier; and sending an authentication challenge message to the wearable security terminal; the authentication challenge message carries the first timestamp, the first random number, the encrypted vehicle identifier, the vehicle-side random number, and the first challenge hash value.

[0370] The first timestamp refers to the current time when the electric two-wheeler generates the authentication challenge message. It is used to identify the generation time of the authentication challenge message and prevent replay attacks. The purpose of the first timestamp is to ensure the timeliness of the authentication message. The receiver can use the timestamp to determine whether the authentication message is within the valid time window. If the timestamp exceeds the valid time window, the authentication message is rejected.

[0371] The first random number refers to a random value generated by the electric two-wheeler to ensure the uniqueness of each authentication challenge message. The purpose of the first random number is to prevent replay attacks. Even if an attacker intercepts the authentication challenge message, because the first random number is different for each authentication, the attacker cannot pass authentication by replaying the old authentication message.

[0372] Vehicle-side random numbers (VSNs) are random values ​​generated by electric two-wheelers to encrypt vehicle identification information, increasing the randomness and unpredictability of the encryption process. The purpose of VSNs is to ensure the security of encrypted vehicle identification; even with the same vehicle identification information, different encrypted vehicle identifications will be generated in different authentication attempts, preventing attackers from inferring the original information by analyzing the encryption results.

[0373] It should be noted that: the first random number is used to derive the first temporary encryption key and the first temporary signature key, serving as the input parameter for the key derivation function; the vehicle-side random number is used as the initialization vector for the encryption algorithm, participating in the encryption process of the electric two-wheeler's identification information. The first random number and the vehicle-side random number have different uses. Furthermore, the first random number needs to be returned by the wearable security terminal in the authentication response message. The electric two-wheeler verifies the authenticity of the authentication response message by checking whether the returned value of the first random number matches the first random number itself; the vehicle-side random number does not need to be returned and is only used during encryption and decryption.

[0374] A key derivation function is a cryptographic function that transforms a master key and optional other input parameters into multiple subkeys. Examples of key derivation functions include the Hash-based Message Authentication Code Key Derivation Function (HKDF), the Password-Based Key Derivation Function 2 (PBKDF2), and the Secure Hash Algorithm Key Derivation Function (SHA-KDF), among others.

[0375] A key derivation function takes the master key as input and combines it with random numbers, salt values, or other parameters. Through cryptographic operations such as hashing and pseudo-random functions, it outputs one or more cryptographically secure subkeys. Key derivation functions are deterministic, meaning that the same input always produces the same output; they are also irreversible, meaning that the master key cannot be derived from the subkeys.

[0376] The master key is the original key used to derive other subkeys and is the core input of the key derivation process. In this embodiment, the authentication key is used as the master key because the authentication key is a unique key generated by the cloud server based on the three-element binding information and has been securely stored in the electric two-wheeler and wearable security terminal. Using the authentication key as the master key can avoid additional key management overhead and ensure that the derived temporary key uniquely corresponds to the three-element binding relationship.

[0377] In this embodiment, the authentication key is used as the master key, the first random number is used as the input parameter, and the key derivation function is input. The key derivation function performs hash operation and pseudo-random function operation on the master key and the first random number, and outputs two subkeys. The first subkey is used as the first temporary encryption key, and the second subkey is used as the first temporary signature key.

[0378] The first temporary encryption key is a temporary key derived through a key derivation function and used to encrypt vehicle identification information. The purpose of the first temporary encryption key is to ensure the security of the encryption process. Because the first temporary encryption key is dynamically derived from a first random number, it is different for each authentication. Even if one temporary key is leaked, it will not affect the security of other authentication processes.

[0379] The first temporary signature key is a temporary key derived through a key derivation function used to generate the challenge hash value. The purpose of the first temporary signature key is to ensure the authenticity and integrity of the challenge hash value. Because the first temporary signature key is dynamically derived from a first random number, attackers cannot forge a valid challenge hash value.

[0380] In this embodiment, although the first temporary encryption key is different for each authentication, the vehicle-side random number is still needed as an initialization vector for encryption. The reasons are as follows:

[0381] Reason 1: To prevent identical plaintext from generating identical ciphertext. If only the first temporary encryption key is used to encrypt the identification information of the electric two-wheeler, in Electronic Codebook (ECB) mode, the same plaintext and the same key will produce the same ciphertext. Attackers could potentially deduce the characteristics or patterns of the vehicle identification information by analyzing the identical ciphertext generated during multiple authentication processes. Introducing a vehicle-side random number as an initialization vector ensures that even with the same plaintext and the same key, the ciphertext generated each time is different, increasing the difficulty of analysis for attackers.

[0382] Reason 2: Preventing pattern analysis attacks. If the same or similar keys are used to encrypt the same plaintext multiple times (vehicle identification information is usually fixed), an attacker could collect the encryption results, analyze the encryption pattern, ciphertext length, ciphertext distribution, and other characteristics to infer partial information about the plaintext or key. Introducing vehicle-side random numbers can break this pattern, making each encryption result random and preventing pattern analysis attacks.

[0383] Reason 3: Enhanced resistance to replay attacks. The vehicle-side random number is generated independently from the first random number, increasing the randomness and unpredictability of the authentication challenge message. Even if an attacker intercepts an authentication challenge message, because the vehicle-side random number is different each time, the attacker cannot pass verification by replaying the old encrypted vehicle identifier, thus enhancing the system's resistance to replay attacks.

[0384] Therefore, in this embodiment, encrypting the identification information of the electric two-wheeler using a first temporary encryption key and a vehicle-side random number is to ensure the confidentiality and security of the vehicle identification information during transmission. The specific encryption process is as follows: the identification information of the electric two-wheeler is used as plaintext, the first temporary encryption key is used as the encryption key, the vehicle-side random number is used as the initialization vector or additional parameter, and a symmetric encryption algorithm (such as the Advanced Encryption Standard AES) is used to encrypt the plaintext to obtain the ciphertext as the encrypted vehicle identification.

[0385] Encrypted vehicle identification refers to the encrypted identification information of an electric two-wheeler, used to securely transmit vehicle identity information in authentication challenge messages. The purpose of encrypted vehicle identification is to prevent theft or tampering of vehicle identification information, ensuring that only a wearable security terminal with the correct authentication key can decrypt and obtain the genuine vehicle identification information.

[0386] In this embodiment, an attacker may intercept the authentication challenge message and modify key information (such as timestamps, random numbers, and encrypted vehicle identifiers), preventing the recipient from correctly verifying the message's origin. For example, if the first challenge hash value is generated solely based on the first temporary signature key, it is unrelated to the specific content of the authentication challenge message (first timestamp, first random number, and encrypted vehicle identifier). An attacker could intercept the authentication challenge message and modify its content, while the first challenge hash value would remain valid and the tampering would go undetected. Furthermore, if the first challenge hash value is generated solely based on the first temporary signature key, an attacker could intercept multiple authentication challenge messages, select content favorable to themselves, and combine them to generate a forged authentication challenge message.

[0387] To avoid the problem of failing to detect tampering, this embodiment uses a first temporary signing key as the signing key, and a first timestamp, a first random number, and an encrypted vehicle identifier as the message content to be signed. Incorporating the first timestamp, the first random number, and the encrypted vehicle identifier into the hash calculation ensures that the first challenge hash value is bound to the message content. Any modification to the message content will cause the hash value verification to fail, thus detecting tampering. It also prevents attackers from extracting parts of multiple messages for combined attacks. Furthermore, the first timestamp is bound to the authentication time context, the first random number is bound to the authentication uniqueness context, and the encrypted vehicle identifier is bound to the authentication identity context. By incorporating this information into the hash calculation, the first challenge hash value can be bound to the complete authentication context, preventing attackers from performing replay or substitution attacks between different authentication sessions.

[0388] In this embodiment, the first temporary signature key, the first timestamp, the first random number, and the encrypted vehicle identifier are concatenated in a preset order to obtain a concatenated string; a hash algorithm (such as the secure hash algorithm SHA-256) is used to perform a hash operation on the concatenated string to obtain a hash value as the first challenge hash value.

[0389] In this embodiment, the authentication challenge message carries a first timestamp, a first random number, an encrypted vehicle identifier, a vehicle-side random number, and a first challenge hash value, which can solve the problems of replay attacks, tampering attacks, and key leakage of authentication messages. Specifically, the first timestamp ensures the timeliness of the message, the first random number ensures the uniqueness of the message, the encrypted vehicle identifier ensures the confidentiality of vehicle identity information, the vehicle-side random number increases the randomness of encryption, and the first challenge hash value ensures the integrity and authenticity of the message.

[0390] Optional, Figure 11 The timing diagram for the two-way encrypted communication between the electric two-wheeler and the wearable security terminal is as follows: Figure 11 As shown, the process includes the following steps: Step S1101, the on-board controller of the electric two-wheeled vehicle obtains the current timestamp as the first timestamp, generates a first random number and a vehicle-side random number, and reads the identification information of the electric two-wheeled vehicle from the local secure storage area;

[0391] In step S1102, the vehicle controller calls the key derivation function, uses the authentication key as the master key, and derives the first temporary encryption key and the first temporary signature key based on the first random number.

[0392] In step S1103, the vehicle controller uses the first temporary encryption key and the vehicle-side random number to encrypt the identification information of the electric two-wheeler to obtain the encrypted vehicle identification.

[0393] Step S1104: The vehicle controller generates a first challenge hash value based on the first temporary signature key, the first timestamp, the first random number, and the encrypted vehicle identifier;

[0394] In step S1105, the vehicle controller packages and encapsulates the first timestamp, the first random number, the encrypted vehicle identifier, the vehicle-side random number, and the first challenge hash value to generate an authentication challenge message, which is then sent to the wearable security terminal.

[0395] Step S1106: The terminal controller of the wearable security terminal receives the authentication challenge message sent by the electric two-wheeler, parses the authentication challenge message, and obtains the first timestamp, the first random number, the encrypted vehicle identifier, the vehicle-side random number, and the first challenge hash value of the authentication challenge message sent by the electric two-wheeler.

[0396] Step S1107: The terminal controller of the wearable security terminal verifies whether the first timestamp is within the preset valid time window. If the first timestamp exceeds the valid time window, it is determined that the authentication challenge message has expired and the authentication fails.

[0397] In step S1108, the terminal controller of the wearable security terminal calls the key derivation function, uses the authentication key as the master key, and derives the first temporary encryption key and the first temporary signature key according to the first random number;

[0398] Step S1109: The terminal controller of the wearable security terminal generates a verification hash value based on the first temporary signature key, the first timestamp, the first random number, and the encrypted vehicle identifier.

[0399] Step S1110: The terminal controller of the wearable security terminal verifies whether the verification hash value is consistent with the first challenge hash value. If they are inconsistent, it is determined that the authentication challenge message has been tampered with and the authentication fails.

[0400] In step S1111, the terminal controller of the wearable security terminal uses the first temporary encryption key and the vehicle-side random number to decrypt the encrypted vehicle identifier and obtain the identification information of the electric two-wheeler.

[0401] In step S1112, the terminal controller of the wearable security terminal obtains the current timestamp as the second timestamp and generates a second random number and a terminal-side random number.

[0402] In step S1113, the terminal controller of the wearable security terminal calls the key derivation function, uses the authentication key as the master key, and derives the second temporary encryption key and the second temporary signature key according to the second random number.

[0403] In step S1114, the terminal controller of the wearable security terminal uses the second temporary encryption key and the terminal-side random number to encrypt the identity of the wearable security terminal and the first random number feedback value to obtain the encrypted terminal identifier.

[0404] Step S1115: The terminal controller of the wearable security terminal generates a second challenge hash value based on the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier;

[0405] In step S1116, the terminal controller of the wearable security terminal packages and encapsulates the second timestamp, the second random number, the encrypted terminal identifier, and the second challenge hash value to generate an authentication response message, which is then sent to the electric two-wheeler.

[0406] The verification hash value refers to the hash value generated by the wearable security terminal based on the first temporary signature key, the first timestamp, the first random number, and the encrypted vehicle identifier. It is used to verify the integrity and authenticity of the authentication challenge message. The verification hash value is compared with the first challenge hash value. If they match, it indicates that the authentication challenge message has not been tampered with during transmission; if they do not match, it indicates that the authentication challenge message has been tampered with or originated from an unauthorized device.

[0407] The second timestamp refers to the current time when the wearable security terminal generates the authentication response message. It is used to identify the generation time of the authentication response message and prevent replay attacks. The purpose of the second timestamp is to ensure the timeliness of the authentication response message. Electric two-wheelers can verify the second timestamp to determine whether the authentication response message was generated within the valid time window.

[0408] The second random number is a random value generated by the wearable security terminal to ensure the uniqueness of each authentication response message. The purpose of the second random number is to prevent replay attacks; even if an attacker intercepts the authentication response message, because the second random number is different for each authentication, the attacker cannot pass verification by replaying old authentication response messages. The second random number is also used to derive the second temporary encryption key and the second temporary signature key.

[0409] Terminal-side random numbers refer to the random values ​​generated by the wearable security terminal to encrypt the wearable security terminal's identity and the first random number returned value. This increases the randomness and unpredictability of the encryption process. The purpose of terminal-side random numbers is to ensure the security of the encrypted terminal identity; even the same wearable security terminal identity will generate different encrypted terminal identities in different authentication attempts, preventing attackers from inferring the original information by analyzing the encryption results.

[0410] The second temporary encryption key refers to the temporary key derived by the wearable security terminal through a key derivation function, used to encrypt the wearable security terminal's identity and the returned value of the first random number. The role of the second temporary encryption key is to ensure the security of the encryption process. Since the second temporary encryption key is dynamically derived based on the second random number, the second temporary encryption key is different for each authentication. Even if one temporary key is leaked, it will not affect the security of other authentication processes.

[0411] The second temporary signature key refers to the temporary key derived by the wearable security terminal through a key derivation function to generate the second challenge hash value. The purpose of the second temporary signature key is to ensure the authenticity and integrity of the second challenge hash value. Since the second temporary signature key is dynamically derived based on the second random number, attackers cannot forge a valid second challenge hash value.

[0412] The encrypted terminal identifier refers to the ciphertext obtained by encrypting the identity identifier of the wearable security terminal and the first random number returned value. It is used to securely transmit the identity information of the wearable security terminal in the authentication response message. The purpose of the encrypted terminal identifier is to prevent the wearable security terminal's identity identifier and the first random number returned value from being stolen or tampered with, ensuring that only the electric two-wheeler with the correct authentication key can decrypt and obtain the real wearable security terminal's identity identifier and the first random number returned value.

[0413] The second challenge hash value refers to the hash value generated by the wearable security terminal based on the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier. It is used to ensure the integrity and authenticity of the authentication response message. The purpose of the second challenge hash value is to prevent the authentication response message from being tampered with; any modification to the message content will cause the hash value verification to fail. The electric two-wheeler verifies the authenticity of the authentication response message by verifying the second challenge hash value.

[0414] This embodiment introduces a timestamp mechanism to ensure the timeliness of authentication messages. Authentication messages intercepted by attackers outside the valid time window will fail verification, effectively preventing replay attacks. A first random number and a vehicle-side random number are generated, while the wearable security terminal generates a second random number and a terminal-side random number. Each authentication uses a different random number, ensuring the uniqueness of the authentication message. This random number mechanism ensures the uniqueness of each authentication message; even if an attacker intercepts multiple authentication messages, they cannot pass verification through replay or combination attacks. A key derivation function is used to derive a temporary encryption key and a temporary signature key based on the authentication key and the random number. The authentication process utilizes a unique temporary key, used only for the current authentication process. This dynamic key updating ensures that even if one temporary key is compromised, it won't affect the security of other authentication processes, mitigating the security risks associated with unchanging keys. A challenge hash value is generated based on the temporary signature key, timestamp, random number, and encrypted identifier. This challenge hash value is bound to the message content; any modification to the message content will cause the hash value verification to fail, achieving message-level digital signatures and ensuring message integrity and authenticity. Attackers cannot tamper with the message content undetected. In summary, by introducing timestamps, random numbers, key derivation functions, temporary keys, and challenge hash values, this approach achieves anti-replay, anti-tampering, and dynamic key updates for authentication messages, constructing a multi-layered security protection system that significantly improves the security and reliability of the authentication process.

[0415] In an exemplary embodiment, the existing verification process of authentication response messages returned by wearable security terminals from electric two-wheeled vehicles has the following drawbacks: First, it cannot effectively verify the timeliness of the authentication response message, allowing attackers to intercept old authentication response messages for replay attacks; second, it cannot effectively verify the integrity of the authentication response message, allowing attackers to tamper with key information within the message; and third, it cannot effectively verify the authenticity of the source of the authentication response message, allowing attackers to forge the message and impersonate a legitimate device. To address these drawbacks, this embodiment parses the authentication response message to obtain key information, verifies the validity of the timestamp, derives a temporary key, generates a challenge hash value for comparison, decrypts the encrypted terminal identifier to obtain identity information, and verifies the matching of the random number return value and the identity identifier. This achieves comprehensive verification of the authentication response message, ensuring its timeliness, integrity, authenticity, and source reliability.

[0416] In some embodiments, the authentication response message is decrypted using an authentication key and then verified, including: parsing the authentication response message to obtain a second timestamp, a second random number, an encrypted terminal identifier, and a second challenge hash value sent by the wearable security terminal; in response to the first timestamp being within a preset valid time window, a second temporary encryption key and a second temporary signature key are derived based on a key derivation function, using the authentication key as the master key; the encrypted terminal identifier is obtained by encrypting the wearable security terminal's identity identifier using the second temporary encryption key, a terminal-side random number, and a first random number feedback value; a third challenge hash value is generated based on the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier; in response to the third challenge hash value being consistent with the first challenge hash value, the encrypted terminal identifier is decrypted using the second temporary encryption key to obtain the wearable security terminal's identity identifier, a terminal-side random number, and a first random number feedback value; in response to the first random number feedback value being consistent with the first random number, and the wearable security terminal's identity identifier matching the locally stored wearable security terminal identity, the authentication response message is determined to have passed verification.

[0417] In this embodiment, the authentication response message carries a second timestamp, a second random number, an encrypted terminal identifier, and a second challenge hash value. The second timestamp, the second random number, the encrypted terminal identifier, and the second challenge hash value have been explained in the above embodiments and will not be repeated here.

[0418] In this embodiment, since the authentication key is a unique key generated by the cloud server based on the ternary binding information and is securely stored in the electric two-wheeler and wearable security terminal, using the authentication key as the master key ensures that the derived temporary key uniquely corresponds to the ternary binding relationship, while avoiding additional key management overhead. Specifically, the authentication key is used as the master key, and the second random number is used as the input parameter to the key derivation function; the key derivation function performs hash operations and pseudo-random function operations on the master key and the second random number, outputting two subkeys; the first subkey is used as the second temporary encryption key, and the second subkey is used as the second temporary signature key.

[0419] In this embodiment, the encrypted terminal identifier refers to the ciphertext obtained by encrypting the identity identifier of the wearable security terminal using a second temporary encryption key, a terminal-side random number, and a first random number return value. The first random number return value refers to the value of the first random number returned by the wearable security terminal in the authentication response message, which is used to verify whether the authentication response message is a correct response to the authentication challenge message.

[0420] Since attackers may intercept authentication response messages and modify key information (such as timestamps, random numbers, and encrypted terminal identifiers), causing electric two-wheelers to be unable to correctly verify the message source, this embodiment generates a third challenge hash value based on the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier to verify the integrity and authenticity of the authentication response message and ensure that the key information in the authentication response message has not been tampered with. The message content is signed by the challenge hash value, and any modification to the message content will cause the hash value verification to fail, thereby detecting the tampering behavior.

[0421] In this embodiment, the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier are concatenated in a preset order to obtain a concatenated string; a hash algorithm (such as the secure hash algorithm SHA-256) is used to perform a hash operation on the concatenated string to obtain a hash value as the third challenge hash value.

[0422] The third challenge hash value being identical to the first challenge hash value indicates that the key information in the authentication response message (second timestamp, second random number, encrypted vehicle identifier) ​​corresponds to the key information in the authentication challenge message (first timestamp, first random number, encrypted vehicle identifier). This means the authentication response message is a correct response to the authentication challenge message, and the message content has not been tampered with during transmission. In this case, the electric two-wheeler uses the second temporary encryption key to decrypt the encrypted terminal identifier to obtain the true identity identifier of the wearable security terminal and the first random number return value, further verifying the authenticity of the wearable security terminal's identity.

[0423] In some embodiments, the method further includes: determining that the authentication response message has not been verified in response to a discrepancy between the third challenge hash value and the first challenge hash value.

[0424] The inconsistency between the third challenge hash value and the first challenge hash value indicates that the key information in the authentication response message does not correspond to the key information in the authentication challenge message, or that the authentication response message has been tampered with during transmission, making the authentication response message unreliable and authentication failure.

[0425] The fact that the first random number returned value is consistent with the first random number indicates that the wearable security terminal has correctly received and processed the authentication challenge message. The authentication response message is a correct response to the authentication challenge message, rather than a replayed old authentication response message or a forged authentication response message.

[0426] In this embodiment, even if the first random number return value is verified, it is still necessary to confirm whether the wearable security terminal that sent the authentication response message is a legitimate device pre-bound by the cloud server. That is, in addition to verifying whether the first random number return value is consistent with the first random number, it is further verified whether the identity of the wearable security terminal matches the locally stored wearable security terminal identity to prevent unbound wearable security terminals from impersonating legitimate devices after obtaining the authentication key through other means.

[0427] In some embodiments, the method further includes: in response to the first random number return value being inconsistent with the first random number, or the identity identifier of the wearable security terminal not matching the locally stored wearable security terminal identity, determining that the authentication response message has not been verified.

[0428] Specifically, if the returned value of the first random number is inconsistent with the first random number, it indicates that the authentication response message is not a response to the current authentication challenge message, but may be a replayed old authentication response message or a forged authentication response message, resulting in authentication failure. If the identity identifier of the wearable security terminal does not match the locally stored identity identifier of the wearable security terminal, it indicates that the wearable security terminal sending the authentication response message is not a legitimate device pre-bound to the cloud server, resulting in authentication failure.

[0429] Optional, Figure 12 Another optional timing diagram for bidirectional encrypted communication between an electric two-wheeler and a wearable security terminal, provided for embodiments of this application, is as follows: Figure 12 As shown, the timing sequence for bidirectional encrypted communication between the electric two-wheeler and the wearable security terminal also includes the following steps:

[0430] In step S1117, the on-board controller of the electric two-wheeler receives the authentication response message returned by the wearable security terminal;

[0431] Step S1118: The vehicle controller parses the authentication response message to obtain the second timestamp, second random number, encrypted terminal identifier, and second challenge hash value of the authentication response message sent by the wearable security terminal;

[0432] Step S1119: The vehicle controller verifies whether the first timestamp is within the preset valid time window;

[0433] Step S1120: If the first timestamp is within the preset valid time window, the vehicle controller calls the key derivation function, uses the authentication key as the master key, and derives the second temporary encryption key and the second temporary signature key according to the second random number.

[0434] Step S1121: The vehicle controller generates a third challenge hash value based on the second temporary signature key, the second timestamp, the second random number, and the encrypted vehicle identifier;

[0435] Step S1122: The vehicle controller verifies whether the third challenge hash value is consistent with the first challenge hash value;

[0436] Step S1123: If the third challenge hash value is consistent with the first challenge hash value, the vehicle controller uses the second temporary encryption key to decrypt the encrypted terminal identifier and obtain the wearable security terminal's identity identifier, terminal-side random number, and first random number return value.

[0437] Step S1124: The vehicle controller verifies whether the returned value of the first random number is consistent with the first random number.

[0438] Step S1125: The vehicle controller verifies whether the identity of the wearable security terminal matches the locally stored identity of the pairable wearable security terminal.

[0439] Step S1126: If the value returned by the first random number is consistent with the first random number, and the identity of the wearable security terminal matches the identity of the pairable wearable security terminal, the vehicle controller determines that the authentication response message has been verified and establishes an encrypted communication link with the wearable security terminal.

[0440] Step S1127: If the hash value of the third challenge is inconsistent with the hash value of the first challenge, the vehicle controller determines that the authentication response message has not been verified and refuses to establish a communication link.

[0441] Step S1128: If the value returned by the first random number is inconsistent with the first random number, the vehicle controller determines that the authentication response message has not been verified and refuses to establish a communication link.

[0442] In step S1129, if the identity of the wearable security terminal does not match the identity of the pairable wearable security terminal, the vehicle controller determines that the authentication response message has not been verified and refuses to establish a communication link.

[0443] This embodiment parses the authentication response message, verifies the validity of the timestamp, derives the temporary key, compares the challenge hash value, decrypts the encrypted terminal identifier, verifies the random number return value and the identity identifier matching, thus achieving comprehensive verification of the authentication response message. This solves the problem of incomplete verification of authentication response messages in related technologies and improves the security, reliability and anti-attack capabilities of the authentication process.

[0444] According to one aspect of the embodiments of this application, a safety protection method for an electric two-wheeled vehicle is provided. Optionally, in this embodiment, the above-described safety protection method for an electric two-wheeled vehicle may be applied, but is not limited to, to applications such as... Figure 1 The hardware environment shown includes an electric two-wheeler 102, a cloud server 104, a wearable security terminal 106, and a preset channel 108.

[0445] The electric two-wheeled vehicle safety protection method of this application embodiment is implemented by the electric two-wheeled vehicle 102. Figure 13 This is a schematic flowchart of an optional safety protection method for an electric two-wheeled vehicle according to an embodiment of this application, as shown below. Figure 13 As shown, the process of this method may include the following steps:

[0446] Step S1302: In response to the detection of a collision event by the electric two-wheeled vehicle, determine the severity of the collision event, and / or, in response to receiving collision detection data sent by the wearable safety terminal, perform secondary detection based on the collision detection data to determine the severity of the collision event;

[0447] Step S1304: In response to the collision severity of the collision event being greater than or equal to a preset rescue trigger threshold, a rescue request is sent to a preset channel; the rescue request includes at least one of the following: the electric two-wheeler's equipment information, rider information, accident type, and location coordinates.

[0448] The electric two-wheeled vehicle safety protection method in this embodiment can be applied to the field of safety protection for two-wheeled vehicles such as two-wheeled motorcycles and electric two-wheeled vehicles, and can be applied to scenarios where rescue is automatically triggered after a rider is involved in a collision.

[0449] Currently, riders primarily rely on independent wearable safety terminals for safety protection. However, existing technologies have the following drawbacks: First, they lack a collision severity assessment mechanism. Collision detection directly triggers rescue efforts, failing to distinguish between minor and severe collisions, leading to wasted rescue resources and user inconvenience. Second, they lack a multi-source data fusion and judgment mechanism, depending solely on a single data source from the vehicle or wearable device, making it impossible to comprehensively assess the authenticity and severity of the collision. Third, rescue response is slow. Traditional manual alarm methods require riders to actively call for help after an accident. If a rider is seriously injured and unable to call for help, the rescue response time is significantly prolonged, missing the golden window for treatment.

[0450] To address the problem of unnecessary rescue triggering caused by the lack of collision severity assessment and secondary detection mechanisms in existing technologies, this embodiment determines the severity of a collision after the electric two-wheeled vehicle detects a collision event, or determines the severity of a collision after receiving collision detection data sent by a wearable safety terminal. A rescue request is only sent when the collision severity is greater than or equal to a preset rescue trigger threshold, thus realizing intelligent rescue triggering decision-making and avoiding unnecessary rescues caused by minor collisions and false triggers.

[0451] In this embodiment, an integrated cloud-based automatic emergency rescue dispatch system is implemented. Upon collision triggering, the system automatically reports the location and accident information, enabling rapid dispatch of rescue resources. This solves the problems of slow emergency response, high rider injury and death rates in existing technologies, which often result in no automatic rescue loop after an accident. The system enables rapid automatic reporting of accident location and information after a collision triggering, and the cloud platform quickly completes rescue dispatch. Compared to traditional manual alarms, the rescue response time is significantly shortened, effectively reducing the waiting time within the golden window for treatment after an accident and significantly reducing rider accident mortality. This also addresses the security and performance issues of existing linkage solutions, such as high communication latency, unreliable data transmission, lack of two-way authentication, and high false trigger rate.

[0452] In this context, the detection of a collision event by an electric two-wheeler signifies that the vehicle's onboard sensor unit has detected abnormal signals such as sudden acceleration changes, attitude changes, or obstacle collisions, classifying it as a suspected collision event. While the detection indicates that the vehicle's onboard sensor unit has sensed the collision, the authenticity and severity of the collision require further confirmation. To avoid unnecessary rescue efforts, this embodiment of the application determines the severity of the collision event after detection. If the severity is greater than or equal to a preset rescue trigger threshold, a rescue request is sent to a preset channel. This collision severity assessment mechanism distinguishes between minor and severe collisions, triggering rescue only in cases of severe collisions. This avoids unnecessary rescues caused by minor collisions, reduces waste of rescue resources and user inconvenience, and improves the intelligence level of rescue decision-making.

[0453] In some embodiments, in response to the detection of a collision event by the electric two-wheeled vehicle, determining the severity of the collision event includes:

[0454] The onboard controller of the electric two-wheeler acquires ...

Claims

1. A safety protection method for an electric two-wheeled vehicle, characterized in that, include: In response to the electric two-wheeler being powered on, the vehicle sensor parameters of the electric two-wheeler are collected, and environmental data of the environment in which the electric two-wheeler is located is received from the cloud server. Based on the vehicle sensor parameters and the environmental data, a collision risk prediction is performed on the electric two-wheeler. In response to the prediction of a collision risk to the electric two-wheeler, at least one safety protection action is performed, and a trigger signal is sent to the wearable safety terminal; The at least one safety protection action includes at least one of: audible and visual warning, instrument display, and vehicle airbag deployment; the trigger signal is used to instruct the wearable safety terminal to perform at least one safety response action, the at least one safety response action includes: performing a body-sensing warning, wearable airbag deployment, and audible and visual warning.

2. The method according to claim 1, characterized in that, The method further includes: In response to a predicted collision risk with the electric two-wheeler, a warning event record is uploaded to the cloud server.

3. The method according to claim 2, characterized in that, The method further includes: The cloud server pushes device status, accident alarms, and rescue progress to preset channels.

4. The method according to claim 1, characterized in that, The method of predicting the collision risk of the electric two-wheeler based on the vehicle sensor parameters and the environmental data includes: Based on the vehicle sensor parameters, determine the first probability of the electric two-wheeler colliding; Based on the vehicle sensor parameters, a second probability of the electric two-wheeler exhibiting an abnormal posture is determined; Based on the environmental data, a third probability of the electric two-wheeler losing control due to environmental factors is determined; The first probability, the second probability, and the third probability are fused together to obtain the predicted collision probability. In response to the predicted collision probability being greater than a preset probability threshold, it is determined that the electric two-wheeler has a collision risk.

5. The method according to claim 4, characterized in that, The vehicle sensor parameters include the relative distance between the electric two-wheeler and the obstacle, and the relative speed between the electric two-wheeler and the obstacle; determining the first probability of a collision between the electric two-wheeler and the obstacle based on the vehicle sensor parameters includes: In response to the relative speed being greater than or equal to 0, the first probability of the electric two-wheeled vehicle colliding with the obstacle is determined to be 0. In response to the relative speed being less than 0, the ratio between the absolute values ​​of the relative distance and the relative speed is determined as the predicted collision duration. A target duration range in which the predicted collision duration is located is determined from multiple duration ranges. The probability value corresponding to the target duration range is determined as the first probability of the electric two-wheeled vehicle colliding.

6. The method according to claim 4, characterized in that, The vehicle sensor parameters include the attitude angle and acceleration of the electric two-wheeler; determining the second probability of the electric two-wheeler exhibiting an abnormal attitude based on the vehicle sensor parameters includes: The ratio between the absolute value of the attitude angle and the attitude angle threshold is determined as the attitude angle anomaly degree. The ratio between the absolute value of the acceleration and the acceleration threshold is determined as the acceleration anomaly degree; The attitude angle anomaly and the acceleration anomaly are fused together to obtain a second probability of the electric two-wheeled vehicle exhibiting an abnormal attitude.

7. The method according to claim 4, characterized in that, The environmental data includes rainfall level, light intensity, and road surface condition; determining the third probability of the electric two-wheeler losing control due to environmental factors based on the environmental data includes: The target rainfall range of the rainfall level is determined from multiple rainfall ranges, and the target illumination range of the illumination intensity is determined from multiple illumination ranges. The third probability of the electric two-wheeled vehicle losing control due to environmental factors is obtained by fusing the first risk probability corresponding to the target rainfall range, the second risk probability corresponding to the target illumination range, and the third risk probability matching the road surface condition.

8. The method according to claim 4, characterized in that, The method further includes: The wearable sensor parameters from the wearable security terminal are received via near-field communication, and a fourth probability of collision is determined based on the wearable sensor parameters of the wearable security terminal. The fourth probability and the predicted collision probability are fused to obtain a new predicted collision probability for the electric two-wheeled vehicle to collide.

9. The method according to any one of claims 1 to 8, characterized in that, The method further includes: A near-field communication connection is established with the wearable security terminal via a wireless transmission protocol; The wearable sensor parameters are periodically sent by the wearable security terminal, and the wearable sensor parameters include at least one of user physiological data and user operation commands; Safety protection decisions are made based on the wearable sensor parameters.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 9.

Images