Remote anti-theft early warning method and system based on vehicle attitude sensor and vehicle

Abstract

This invention discloses a remote anti-theft warning method, system, and vehicle based on vehicle attitude sensors. The method includes constructing a vehicle attitude reference model, receiving a remote locking command, and activating an anti-theft monitoring mode; acquiring the vehicle's angular velocity and linear acceleration data and performing filtering and fusion processing to obtain the vehicle's dynamic attitude angles and processed triaxial linear accelerations; performing dual-layer parallel anomaly detection on the dynamic attitude angles and triaxial accelerations to identify abnormal features related to tire removal and vehicle lifting attitude deviation, respectively, to obtain anomaly detection results; determining the type of vehicle anomaly event through collaborative event judgment using a finite state machine based on the anomaly detection results; and generating an alarm data packet containing event type information and vehicle attitude information when the vehicle anomaly event type meets the alarm conditions, and pushing a warning notification. This invention achieves end-to-end anti-theft control from vehicle attitude monitoring to precise warnings via the owner's app.

Description

Technical Field

[0001] This invention relates to the field of automotive electronics technology, and in particular to a remote anti-theft warning method, system, and vehicle based on vehicle attitude sensors. Background Technology

[0002] Current vehicle anti-theft systems primarily use vibration and tilt sensors for monitoring. The overall anti-theft logic revolves around a single vibration intensity threshold or static tilt angle threshold, and is managed uniformly by the vehicle control module. When abnormal vibration occurs in the vehicle, the vibration sensor detects the vibration and sends an alarm to the owner.

[0003] Existing vehicle anti-theft technologies have the following drawbacks: First, traditional vibration sensor anti-theft solutions have fixed thresholds, making them susceptible to environmental interference such as thunder, heavy vehicles passing by, and nearby vehicles closing their doors, leading to false alarms. They also miss alarms because thieves' slow movements may not trigger the vibration threshold, resulting in a trade-off between monitoring sensitivity and anti-interference capabilities. Second, existing tilt sensor anti-theft solutions only rely on static threshold judgments and cannot identify specific micro-motion-attitude time-domain sequences when a vehicle is illegally lifted or its tires are removed. They are insensitive to "slow" theft and have delayed warning times. Third, existing anti-theft systems only inform the owner that "there is an anomaly in the vehicle," without providing precise information such as the type of abnormal event, changes in vehicle posture, and the location of the event, resulting in low emergency response efficiency for the owner. Fourth, some anti-theft solutions require the addition of dedicated monitoring sensors and control units, leading to high hardware modification costs, poor compatibility with existing vehicle intelligent network hardware architectures, and difficulty in rapid deployment in mass-produced models. Summary of the Invention

[0004] To address the aforementioned issues, this invention proposes a remote anti-theft warning method, system, and vehicle based on vehicle attitude sensors. This invention achieves end-to-end anti-theft control, from vehicle attitude monitoring to precise warnings via the owner's app.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a remote anti-theft warning method based on a vehicle attitude sensor, comprising the following steps: Construct a vehicle attitude reference model, which is used to characterize the reference attitude range of the vehicle under normal parking conditions; Receives a remote vehicle locking command and activates the anti-theft monitoring mode; The vehicle's angular velocity and linear acceleration data are acquired, and the angular velocity and linear acceleration data are filtered and fused to obtain the vehicle's dynamic attitude angles and the processed triaxial linear acceleration. A two-layer parallel anomaly detection was performed on the dynamic attitude angle and triaxial linear acceleration to identify the abnormal features of vehicle tire removal and the vehicle lifting attitude deviation features, and the anomaly detection results were obtained. Based on the anomaly detection results, a finite state machine is used to make cooperative event judgments to determine the type of vehicle anomaly event; When a vehicle abnormal event type meets the alarm conditions, an alarm data packet containing event type information and vehicle attitude information is generated, and a warning notification is pushed.

[0006] A further technical solution involves the following steps to construct a vehicle attitude reference model: During the vehicle factory calibration phase, static reference attitude angles are obtained, including roll angle and pitch angle. By learning the posture data corresponding to the car owner's daily parking positions, a normal posture tolerance range corresponding to the parking position is established. The static reference attitude angle and the normal attitude tolerance range together constitute the vehicle attitude reference model.

[0007] A further technical solution involves filtering and fusing the angular velocity and linear acceleration data, specifically as follows: Low-pass filtering is applied to angular velocity and linear acceleration data to suppress high-frequency noise, resulting in filtered data. The filtered data is then fused using a Kalman filter algorithm to calculate dynamic attitude angles, angular velocities, and triaxial linear accelerations after gravity compensation in real time.

[0008] A further technical solution is that the dual-layer parallel anomaly detection includes high-frequency micro-motion mode detection and low-frequency attitude trajectory tracking; wherein, the high-frequency micro-motion mode detection is as follows: frequency domain transformation is performed on the three-axis linear acceleration to extract the envelope shape of the vibration energy in the preset frequency band on the time axis; the envelope shape is correlated with the pre-stored tire disassembly vibration template, and when the matching degree exceeds the preset threshold, it is determined that the tire disassembly anomaly feature has been detected.

[0009] A further technical solution is that the low-frequency attitude trajectory tracking is as follows: continuously monitor the change of dynamic attitude angle at a preset update rate; when the dynamic attitude angle changes continuously in one direction at a rate lower than the preset angular rate, and the cumulative change exceeds the preset angle value and continues for a preset duration, it is determined that the vehicle lifting attitude deviation feature has been detected.

[0010] In a further technical solution, the finite state machine includes an idle state, a monitoring state, a suspected lifting state, a suspected theft state, and an alarm state.

[0011] Secondly, the present invention provides a remote anti-theft warning system based on a vehicle attitude sensor, comprising the following modules: The baseline establishment module is configured to: construct a vehicle attitude baseline model, which is used to characterize the baseline attitude range of the vehicle in a normal parking state; The anti-theft mode activation module is configured to: receive a remote vehicle locking command and activate the anti-theft monitoring mode; The data acquisition and fusion module is configured to: acquire the vehicle's angular velocity data and linear acceleration data, filter and fuse the angular velocity data and linear acceleration data to obtain the vehicle's dynamic attitude angle and the processed triaxial linear acceleration; The anomaly detection module is configured to perform dual-layer parallel anomaly detection on dynamic attitude angles and triaxial linear acceleration, respectively identifying abnormal features of vehicle tire removal and vehicle lifting attitude deviation, and obtaining anomaly detection results. The event decision module is configured to: determine the type of vehicle abnormal event by performing collaborative event decision through a finite state machine based on the anomaly detection results; The early warning push module is configured to generate an alarm data packet containing event type information and vehicle attitude information when the abnormal vehicle event type meets the alarm conditions, and push an early warning notification.

[0012] Thirdly, the present invention provides a vehicle including a memory and a processor, and computer instructions stored in the memory and running on the processor, wherein the computer instructions, when executed by the processor, perform the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of the first aspects.

[0013] Fourthly, the present invention provides a computer-readable storage medium for storing computer instructions, which, when executed by a processor, perform the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of the first aspects.

[0014] Fifthly, the present invention provides a computer program product comprising executable instructions stored in a computer-readable storage medium; wherein, when the vehicle's processor reads and executes the executable instructions from the computer-readable storage medium, it completes the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of the first aspects.

[0015] Compared with the prior art, the beneficial effects of the present invention are as follows: This invention employs a dual-layer parallel anomaly detection mechanism to separately identify the characteristic vibration envelope of tire removal and the slow deviation of vehicle lifting posture. It can trigger an early warning when the vehicle tilt angle is approximately 2° to 3°, responding earlier than traditional vibration intensity threshold schemes. This enables proactive identification of "slow" theft behaviors such as jack lifting and tire removal, overcoming the shortcomings of existing technologies that are insensitive to such behaviors and have delayed warning timing. Furthermore, by using a finite state machine to correlate and collaboratively decide the dual-layer detection results, the vehicle lifting posture deviation characteristics, jacking initiation pulse characteristics, and tire removal anomaly characteristics must corroborate each other temporally. This effectively filters out false triggers caused by accidental environmental vibrations such as thunder, heavy vehicles passing by, or nearby vehicles closing doors, solving the technical problems of high false alarm and high false negative rates caused by the single monitoring logic in existing technologies. Attached Figure Description

[0016] The accompanying drawings, which form part of this invention, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an improper limitation of the invention.

[0017] Figure 1 This is a schematic diagram of the overall process of the remote anti-theft early warning method based on vehicle attitude sensor of the present invention. Detailed Implementation

[0018] The embodiments of this application will now be described in more detail with reference to the accompanying drawings. While some embodiments of this application are shown in the drawings, it should be understood that this application can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of this application. It should be understood that the drawings and embodiments of this application are for illustrative purposes only and are not intended to limit the scope of protection of this application. The reference numerals in the drawings correspond one-to-one with the textual descriptions in the specification. The drawings only show the necessary structures related to the present invention; components not shown in the drawings do not necessarily mean they do not exist in actual products. The various embodiments of this application can be combined with each other, and the technical features in each embodiment can be referenced and supplemented by each other, provided that the technical solutions do not contradict each other. The embodiments of this application are mainly described in detail for the scenario of expanding the seating space in small SUV non-pure electric vehicles. However, it should be understood that the technical solutions of this application are also applicable to other vehicle models constrained by space limitations such as vehicle size, wheelbase, and fuel tank layout, including but not limited to compact cars and MPVs.

[0019] The invention will now be described in further detail with reference to the accompanying drawings. This invention provides a remote anti-theft early warning method based on vehicle attitude sensors. Its core improvement lies in: using the vehicle's existing six-axis inertial measurement unit as the sensing basis, and through dual-layer parallel anomaly detection and five-state finite state machine collaborative decision-making, it achieves advanced identification of "slow" theft behaviors such as jack lifting and tire removal. This solves the technical problems of high false alarm rate, high false alarm rate, and low intelligence value of alarm information in existing single-threshold judgment schemes.

[0020] Example 1 Existing anti-theft solutions have the following drawbacks: First, vibration sensor anti-theft solutions have fixed thresholds, making them susceptible to environmental interference such as thunder, heavy vehicles passing by, and nearby vehicles closing their doors, resulting in false alarms. They also fail to detect when thieves operate slowly and fail to trigger the vibration threshold, making it difficult to balance monitoring sensitivity and anti-interference capabilities. Second, tilt sensor anti-theft solutions only rely on static threshold judgments and cannot identify the specific micro-motion-attitude time-domain sequences when a vehicle is illegally lifted or its tires are removed. They are insensitive to "slow" theft and have delayed warning times. Third, existing anti-theft systems only inform the owner that "there is an anomaly in the vehicle," without providing precise information such as the type of abnormal event, changes in vehicle posture, and the location of the event, resulting in low emergency response efficiency for the owner. Fourth, some anti-theft solutions require the addition of dedicated monitoring sensors and control units, leading to high hardware modification costs, poor compatibility with existing vehicle intelligent network hardware architectures, and difficulty in rapid deployment in mass-produced models.

[0021] Based on this, this embodiment proposes a remote anti-theft early warning method based on vehicle attitude sensors, such as... Figure 1 As shown, this method relies on the vehicle's existing hardware architecture. The hardware involved includes: a six-axis inertial measurement unit, a body domain controller, a body control module, a remote communication terminal, an in-vehicle camera, an in-vehicle microphone, an in-vehicle bus, a cloud server, a vehicle owner's mobile terminal, and vehicle speakers and lights. The components are interconnected through the in-vehicle bus or a dedicated interface, without the need for additional independent control units.

[0022] The six-axis inertial measurement unit is installed in the central tunnel of the vehicle near the center of gravity. It includes a three-axis accelerometer and a three-axis gyroscope. The accelerometer range is ±8g and the gyroscope range is ±500° / s. It is connected to the vehicle domain controller 2 through a dedicated serial peripheral interface or a high-speed controller LAN channel to collect the vehicle's angular velocity and linear acceleration data.

[0023] The vehicle domain controller is the main control unit of the system, integrating a sensor signal processing module and an anti-theft strategy module. The sensor signal processing module is responsible for filtering, fusing, and attitude calculation of the raw data from the six-axis inertial measurement unit; the anti-theft strategy module is responsible for anomaly detection, event judgment, and alarm command generation.

[0024] The body control module communicates with the body domain controller via the vehicle bus. After receiving the remote locking command and completing the vehicle locking action, it broadcasts an anti-theft activation signal to the vehicle bus, triggering the anti-theft function of the body domain controller.

[0025] The remote communication terminal acts as a communication gateway between the vehicle and the cloud. It connects to the vehicle domain controller via Ethernet or high-speed controller LAN, uploads alarm data packets to the cloud server via 4G / 5G cellular network, and can receive quick operation commands from the vehicle owner's mobile terminal to control the operation of the vehicle camera, vehicle microphone, vehicle speakers, and vehicle lights.

[0026] The cloud server receives alarm data uploaded by the remote communication terminal, parses it, and pushes warning notifications to the vehicle owner's mobile terminal. It also stores vehicle attitude reference data and anti-theft event logs. The vehicle owner's mobile terminal is used to send remote locking commands, receive warning notifications, and perform quick operations.

[0027] Based on the aforementioned hardware, such as Figure 1 As shown, the remote anti-theft warning method based on a vehicle attitude sensor provided in this embodiment specifically includes the following steps: S1: Construct a vehicle attitude reference model, which is used to characterize the reference attitude range of the vehicle under normal parking conditions.

[0028] The purpose of step S1 is to establish a personalized attitude judgment benchmark for subsequent anomaly detection, solve the problem that the existing technology uses a fixed threshold that does not match the actual parking environment, and enable the system to maintain accurate anomaly recognition capability in different parking scenarios.

[0029] The steps for constructing a vehicle attitude baseline model are as follows: During the vehicle factory calibration phase, static reference attitude angles are obtained, including roll angle and pitch angle. By learning the attitude data corresponding to the owner's daily parking positions, a normal attitude tolerance range corresponding to the parking position is established. The static reference attitude angles and the normal attitude tolerance range together constitute the vehicle attitude reference model.

[0030] Specifically: Regarding factory static calibration: Before the vehicle leaves the factory or after maintenance, the body domain controller controls the six-axis inertial measurement unit to complete the static calibration on the horizontal calibration platform, and records the static reference attitude angles output by the IMU at this time. The static reference attitude angles include roll angle and pitch angle.

[0031] The calibration process requires the vehicle to be parked on a horizontal platform with an inclination of no more than 0.1°. The six-axis inertial measurement unit continuously collects static data for no less than 30 seconds, takes the average value as the factory reference value, and writes the reference value into the non-volatile memory of the vehicle body domain controller for long-term storage.

[0032] Regarding learning parking locations in daily use: After a vehicle is delivered to its owner, the cloud server continuously collects IMU static pose data for the first 5 minutes after the owner locks the car at frequently used parking locations (such as home garages and company parking lots). This data, combined with geographic location information, is used to cluster parking locations and gradually establish a "normal pose tolerance range" for each location. The normal pose tolerance range can be set as an interval ±0.5° above and below the average historical pose value of that location, and is dynamically adjusted as learning samples accumulate to adapt to parking scenarios with inherent tilt angles, such as ramp parking spaces.

[0033] Regarding the construction of the vehicle attitude reference model: the static reference attitude angle obtained from the factory static calibration and the normal attitude tolerance range learned from daily parking positions are used together to form the vehicle attitude reference model. When the system determines that the current parking position of the vehicle belongs to a learned location, the normal attitude tolerance range corresponding to that location is used as the anomaly detection reference first; when the parking position is an unlearned location, it reverts to the factory static reference attitude angle superimposed with the default tolerance range (e.g., ±1.0°) as the reference, ensuring that the system can work normally in any parking scenario.

[0034] S2: Receives remote vehicle locking command and activates anti-theft monitoring mode.

[0035] The core purpose of step S2 is to link the owner's remote locking operation with the activation of the anti-theft system, ensuring that the anti-theft monitoring starts immediately after the vehicle is physically locked, while controlling the system power consumption within the range required by the vehicle power management specifications through intermittent operation to ensure monitoring sensitivity.

[0036] Regarding the transmission chain of the vehicle locking command: The vehicle owner sends a remote locking command through an app on their mobile device. This command is uploaded to the cloud server via a 4G / 5G cellular network and then transmitted to the vehicle's remote communication terminal through a secure channel. After receiving the command, the remote communication terminal forwards the locking request to the body control module via the vehicle bus. The body control module then performs physical locking actions such as locking the doors and closing the windows to complete the vehicle locking process.

[0037] Regarding the broadcast mechanism of the anti-theft activation signal: After the body control module completes all locking actions and confirms that the vehicle is in a locked state, it broadcasts an "enter high-level anti-theft state" activation signal to the vehicle bus. The body domain controller continuously listens for bus messages on the vehicle bus. Upon receiving the aforementioned anti-theft activation signal, it immediately wakes up the six-axis inertial measurement unit and puts it into working mode, thus officially starting anti-theft monitoring.

[0038] Regarding the two phases of the intermittent operation mode: the anti-theft monitoring mode adopts an intermittent operation mode of "initial full-speed phase + periodic wake-up phase". Initial full-speed phase: the initial phase after anti-theft activation (e.g., the first 5 minutes), the six-axis inertial measurement unit runs continuously at full speed with the highest sampling frequency, and the anti-theft strategy module simultaneously performs dual-layer parallel anomaly detection to monitor the vehicle status with the highest sensitivity, preventing thieves from immediately committing theft when the vehicle is parked.

[0039] Periodic wake-up phase: If no abnormalities are detected during the initial full-speed phase, the system automatically switches to periodic wake-up mode. The six-axis inertial measurement unit (IMU) intermittently collects data according to a preset cycle (e.g., working at full speed for 5 seconds every 30 seconds). During the remaining time, the IMU enters a low-power standby state. Regarding power consumption control: Through the aforementioned intermittent operation, the increase in vehicle static current caused by the anti-theft system can be controlled within 5mA, and the vehicle static current is below 50mA, complying with vehicle power management specifications and preventing battery depletion due to prolonged vehicle parking.

[0040] S3: Acquire the vehicle's angular velocity and linear acceleration data, filter and fuse the angular velocity and linear acceleration data to obtain the vehicle's dynamic attitude angles and the processed triaxial linear acceleration.

[0041] The core purpose of step S3 is to process the raw angular velocity data and linear acceleration data output by the six-axis inertial measurement unit 1 through multi-level filtering and fusion to calculate high-precision dynamic attitude angles, angular velocities, and three-axis linear accelerations after gravity compensation, providing a reliable and low-noise data foundation for subsequent dual-layer parallel anomaly detection.

[0042] The six-axis inertial measurement unit (IMU) acquires angular velocity and linear acceleration data at a sampling frequency of no less than 50Hz, with a preferred sampling frequency of 100Hz. This ensures that the high-frequency micro-motion mode detection channel can capture tire disassembly characteristic vibration signals within the 5Hz to 15Hz frequency band, while simultaneously meeting the real-time monitoring requirements of the low-frequency attitude trajectory tracking channel for attitude angle change rate. The acquired raw data is transmitted in real-time to the sensor signal processing module of the vehicle domain controller via a dedicated serial peripheral interface or a high-speed controller area network (Controller Area Network) channel, with a transmission delay controllable to within 1ms.

[0043] The sensor signal processing module first performs low-pass filtering on the raw angular velocity and linear acceleration data to remove high-frequency interference components unrelated to theft, such as the sensor's own electrical noise and high-frequency structural vibrations of the vehicle body. The low-pass filter can be a second-order Butterworth low-pass digital filter with a cutoff frequency of 20Hz. This filter has a maximum flat amplitude-frequency response at the cutoff frequency, effectively suppressing high-frequency noise above 20Hz while maintaining good amplitude-frequency transmission characteristics in the 5Hz to 15Hz tire removal characteristic vibration frequency band, ensuring the accuracy of subsequent feature extraction.

[0044] The low-pass filtered data enters the Kalman filter algorithm module. This module uses the gravity vector measured by the accelerometer and the angle change obtained by the gyroscope integration as state variables to establish a dynamic state equation for the vehicle's attitude. Through a predictive-update iterative loop, it fuses the two types of sensor data in real time, outputting the following three types of fusion results: 1) dynamic attitude angles, including roll and pitch angles, reflecting the vehicle's current spatial attitude; 2) angular velocity, reflecting the rate of change of the vehicle's attitude; and 3) triaxial linear acceleration after gravity compensation, i.e., the actual vehicle acceleration obtained by subtracting the gravity component from the original accelerometer output, used for feature extraction in the high-frequency micro-motion mode detection channel. The attitude calculation accuracy can reach 0.1°, meeting the accuracy requirement of the low-frequency attitude trajectory tracking channel for a cumulative change detection threshold of 0.8°.

[0045] S4: Perform dual-layer parallel anomaly detection on dynamic attitude angles and triaxial linear acceleration to identify abnormal features of vehicle tire removal and vehicle lifting attitude deviation, and obtain anomaly detection results.

[0046] The design concept of step S4 is that tire removal behavior and vehicle lifting behavior exhibit distinct signal characteristics in the time and frequency domains. The former is manifested as a high-frequency periodic vibration envelope within a specific frequency band, while the latter is manifested as a low-speed unidirectional cumulative offset of the attitude angle. Therefore, two independent detection channels are used to specifically identify the above two types of characteristics. The two channels are executed in parallel without interfering with each other, and the detection results are output to the finite state machine for collaborative decision-making.

[0047] The dual-layer parallel anomaly detection includes high-frequency micro-motion pattern detection and low-frequency attitude trajectory tracking.

[0048] The high-frequency micro-motion mode detection involves: performing a frequency domain transformation on the triaxial linear acceleration to extract the envelope shape of the vibration energy on the time axis within a preset frequency band; matching the envelope shape with a pre-stored tire disassembly vibration template; and determining that an abnormal tire disassembly feature has been detected when the matching degree exceeds a preset threshold. The specific steps are as follows: First, frequency domain transformation. The frequency domain transformation of the gravity-compensated triaxial linear acceleration data can be performed using short-time Fourier transform (STFT) or wavelet analysis to convert the time-domain acceleration signal into a time-frequency domain representation, obtaining the two-dimensional time-frequency distribution of vibration energy across each frequency band. For STFT, a Hanning window can be used, with a window length of 256 sampling points and a frame shift of 64 sampling points; for wavelet analysis, the Daubechies wavelet basis can be used, with a decomposition level of up to 5 levels.

[0049] Second, envelope extraction. Extract the envelope shape of vibration energy on the time axis within a preset frequency band (which can be from 5Hz to 15Hz) from the time-frequency domain distribution. The envelope shape reflects the fluctuation and change law of vibration energy within the frequency band over time and is described as an energy-time curve.

[0050] Third, template matching. The extracted envelope shape is correlated with the tire removal vibration template pre-stored in the non-volatile memory of the vehicle body domain controller. The tire removal vibration template was obtained during the project development phase through laboratory bench simulation and real vehicle data acquisition. It describes the typical envelope shape (exhibiting 5 to 8 "bumps" with increasing amplitude) of the vibration energy in the 8Hz to 12Hz frequency band when the wheel nuts are loosened. The feature vector is described by wavelet coefficients or specific frequency band energy ratios. Correlation matching uses a normalized cross-correlation algorithm to calculate the similarity score between the measured envelope and the template.

[0051] Fourth, feature confirmation. When the similarity score exceeds a preset threshold, it is determined that an abnormal tire disassembly feature has been detected, and a tire disassembly abnormal feature flag is output to the finite state machine.

[0052] Low-frequency attitude trajectory tracking involves continuously monitoring changes in dynamic attitude angles at a preset update rate. When the dynamic attitude angle changes unidirectionally at a rate lower than a preset angular rate, and the cumulative change exceeds a preset angle value for a preset duration, a vehicle lifting attitude deviation characteristic is detected. The specific steps are as follows: First, update rate setting. The dynamic attitude angles (roll and pitch angles) of the Kalman filter fusion output are continuously sampled at a preset update rate (e.g., 1Hz) to establish an attitude angle time series for subsequent trend analysis.

[0053] Second, unidirectional change judgment. Real-time analysis of the attitude angle time series is performed to determine whether the attitude angle is continuously changing unidirectionally at a rate lower than a preset angular rate (e.g., 2° / s). The "continuous unidirectional change" requires that the direction of attitude angle change remains consistent across multiple consecutive sampling points (either all increasing or all decreasing), excluding random bidirectional fluctuations caused by road vibration or wind load.

[0054] Third, the cumulative change threshold judgment. Based on the confirmation that the attitude angle is continuously changing in one direction, the cumulative change of the attitude angle from the start of the change to the current time is calculated. When the cumulative change exceeds the preset angle value (e.g., 0.8°), the cumulative change threshold flag is triggered.

[0055] Fourth, duration threshold judgment. Based on the cumulative change exceeding the threshold, it is further judged whether the above unidirectional change process has lasted for a preset duration (e.g., 3 seconds) or more, in order to exclude short-term attitude deviation caused by brief disturbances of the people in the vehicle.

[0056] Fifth, feature confirmation. When the three conditions of unidirectional change judgment, cumulative change exceeding the threshold, and duration exceeding the threshold are simultaneously met, it is determined that a vehicle lifting posture deviation feature has been detected, and a vehicle lifting posture deviation feature flag is output to the finite state machine. The two detection channels execute in parallel within the anti-theft strategy module 22, each maintaining an independent data buffer and feature judgment logic, without interfering with each other. The feature output of one channel does not affect the normal operation of the other channel.

[0057] S5: Based on the anomaly detection results, determine the type of vehicle anomaly event by using a finite state machine to make a collaborative event decision.

[0058] The core purpose of step S5 is to perform time-series correlation analysis on the feature flags output by the dual-layer parallel anomaly detection channel. By requiring multiple anomaly features to corroborate each other in time, false triggers caused by single sporadic anomalies are eliminated, thereby accurately identifying compound theft events such as "illegal lifting" and "lifting and removing tires".

[0059] The anti-theft strategy module implements a five-state finite state machine. The five states are defined as follows: the idle state (IDLE) is the inactive state of the anti-theft system, and the six-axis inertial measurement unit is in a low-power standby mode and does not execute any detection logic. The monitoring status (MONITOR) indicates that the anti-theft system is activated and operating normally. The dual-layer parallel anomaly detection channel is continuously running, and the finite state machine continuously receives the characteristic flag outputs of the two channels. Suspected Lift State (SUSPECT_LIFT) is an intermediate state in which the system detects that the vehicle may be being lifted. In this state, the system maintains high sensitivity monitoring and starts timing. The suspected theft status (SUSPECT_THEFT) is a high-confidence abnormal state in which the system detects a combination of lifting and tire removal behavior; The alarm state (ALARM) is the final state in which the system confirms an abnormal event and generates an alarm command. The transition conditions between each state are as follows: After the system is powered on, the finite state machine enters the idle state; after receiving the anti-theft activation signal, the finite state machine switches from the idle state to the monitoring state and begins to continuously receive the dual-layer detection results.

[0060] In monitoring mode, when both the vehicle lifting posture deviation feature flag and the lifting start pulse feature flag are detected simultaneously, the system switches to a suspected lifting state and starts timing, with a timing duration of up to 10 seconds. The lifting start pulse feature is a single-peak pulse generated along the Z-axis at the moment the jack tray contacts the vehicle's bottom edge beam. This pulse duration can be 80ms to 120ms, and the amplitude can be 0.3g to 0.7g. The feature vector consists of the pulse width and peak value, obtained during the project development phase through laboratory bench simulation and real-vehicle data acquisition. It is pre-stored in the non-volatile memory of the vehicle domain controller 2 and identified in real time by the high-frequency micro-motion mode detection channel. In the suspected lifting state, if the same-side tire removal abnormality feature flag is detected, the system immediately switches to a suspected theft state without waiting for the timing to end. If no tire removal abnormality feature is detected after 10 seconds of timing, the system directly switches to the alarm state. Upon entering the suspected theft state, the finite state machine unconditionally switches to the alarm state and generates an alarm command.

[0061] Through the above state transition logic, the system requires that the vehicle lifting posture deviation characteristics, lifting start pulse characteristics, and tire removal abnormal characteristics corroborate each other in time sequence. Any single occasional abnormality (such as short-term posture fluctuation caused only by road vibration, or only a single impact pulse without continuous posture deviation) cannot trigger the state machine to enter the alarm state on its own, thereby effectively eliminating false triggering.

[0062] The timing during suspected lifting is implemented by a software timer within the anti-theft strategy module, with a timing accuracy of 100ms. During the timing period, a dual-layer parallel anomaly detection channel runs continuously, and a finite state machine continuously polls the detection results. If an abnormal feature of tire removal on the same side is detected during the timing process, the timing immediately stops and a state transition is triggered; if the timing expires, an alarm is triggered regardless of whether an abnormal feature of tire removal is detected, to prevent thieves from taking other destructive actions after lifting the vehicle without immediately removing the tires.

[0063] S6: When the abnormal vehicle event type meets the alarm conditions, generate an alarm data packet containing event type information and vehicle attitude information, and push a warning notification.

[0064] The core purpose of step S6 is to transform the alarm command generated by the finite state machine decision into an alarm data packet containing multi-dimensional and accurate information, and to push it securely and quickly to the car owner's mobile terminal through the complete vehicle-cloud-APP transmission link, so that the car owner can grasp the abnormal situation of the vehicle and take countermeasures as soon as possible.

[0065] When a vehicle abnormal event type meets the alarm conditions, the vehicle domain controller generates an alarm data packet, which specifically includes the following four types of information: First, there is event type information, which identifies the type of theft event that triggered the alarm, such as "suspected illegal lifting" or "suspected lifting and tire removal," represented by an enumeration code. Second, vehicle attitude information, including real-time roll and pitch angle values ​​at the time of alarm triggering, recorded with a resolution of 0.1°, intuitively reflecting the current tilt of the vehicle; Third, the timestamp of the event, using UTC time format, with an accuracy of milliseconds, is used to record the exact time of the alarm event, providing a time basis for subsequent accident identification; Fourth, the event's estimated location information is provided by a remote communication terminal based on the most recent valid global navigation satellite system positioning data or base station positioning data, marking the geographical coordinates of where the vehicle is parked, to help vehicle owners and the police quickly locate the vehicle.

[0066] After the alarm data packet is generated by the vehicle domain controller, it is sent to the remote communication terminal via the vehicle bus as a high-priority CAN message. The communication baud rate of the vehicle bus meets the real-time transmission requirements of vehicle signals, and the transmission delay can be controlled within 10ms. After receiving the alarm data packet, the remote communication terminal 4 uses an encrypted communication protocol (such as MQTT over TLS) to upload the encrypted data to the cloud server via the 4G / 5G cellular network. After receiving and decrypting the alarm data packet, the cloud server queries the owner account information bound to the vehicle VIN code and sends a high-priority push notification to the owner's mobile terminal through the push service to ensure that the warning information reaches the owner in a timely manner.

[0067] After receiving the warning notification, the vehicle owner's mobile terminal displays the vehicle's tilt status graphically. The current roll and pitch angle values ​​are presented intuitively through dynamic rendering of the vehicle's 3D model tilt angle. At the same time, detailed information such as event type, occurrence time, and estimated location of the event are displayed in text form.

[0068] The app interface also provides three quick operation entry points: one-click photo taking, horn honking and light flashing, and voice monitoring. When the car owner clicks "one-click photo taking," the in-vehicle camera can be remotely activated to take pictures of the vehicle's surroundings and send back the images; clicking "horn honking and light flashing" can control the vehicle's speakers and lights to issue warnings; clicking "voice monitoring" can use the in-vehicle microphone to listen to the ambient sounds around the vehicle in real time. All of the above quick operation commands are sent from the cloud server to the remote communication terminal for execution.

[0069] Example 2 This embodiment provides a remote anti-theft warning system based on a vehicle attitude sensor, including the following modules: The baseline establishment module is configured to: construct a vehicle attitude baseline model, which is used to characterize the baseline attitude range of the vehicle in a normal parking state; The anti-theft mode activation module is configured to: receive a remote vehicle locking command and activate the anti-theft monitoring mode; The data acquisition and fusion module is configured to: acquire the vehicle's angular velocity data and linear acceleration data, filter and fuse the angular velocity data and linear acceleration data to obtain the vehicle's dynamic attitude angle and the processed triaxial linear acceleration; The anomaly detection module is configured to perform dual-layer parallel anomaly detection on dynamic attitude angles and triaxial linear acceleration, respectively identifying abnormal features of vehicle tire removal and vehicle lifting attitude deviation, and obtaining anomaly detection results. The event decision module is configured to: determine the type of vehicle abnormal event by performing collaborative event decision through a finite state machine based on the anomaly detection results; The early warning push module is configured to generate an alarm data packet containing event type information and vehicle attitude information when the abnormal vehicle event type meets the alarm conditions, and push an early warning notification.

[0070] The system execution method steps provided in this embodiment are as follows: S1: Construct a vehicle attitude reference model, which is used to characterize the reference attitude range of the vehicle under normal parking conditions.

[0071] The purpose of step S1 is to establish a personalized attitude judgment benchmark for subsequent anomaly detection, solve the problem that the existing technology uses a fixed threshold that does not match the actual parking environment, and enable the system to maintain accurate anomaly recognition capability in different parking scenarios.

[0072] The steps for constructing a vehicle attitude baseline model are as follows: During the vehicle factory calibration phase, static reference attitude angles are obtained, including roll angle and pitch angle. By learning the attitude data corresponding to the owner's daily parking positions, a normal attitude tolerance range corresponding to the parking position is established. The static reference attitude angles and the normal attitude tolerance range together constitute the vehicle attitude reference model.

[0073] Specifically: Regarding factory static calibration: Before the vehicle leaves the factory or after maintenance, the body domain controller controls the six-axis inertial measurement unit to complete the static calibration on the horizontal calibration platform, and records the static reference attitude angles output by the IMU at this time. The static reference attitude angles include roll angle and pitch angle.

[0074] The calibration process requires the vehicle to be parked on a horizontal platform with an inclination of no more than 0.1°. The six-axis inertial measurement unit continuously collects static data for no less than 30 seconds, takes the average value as the factory reference value, and writes the reference value into the non-volatile memory of the vehicle body domain controller for long-term storage.

[0075] Regarding learning parking locations in daily use: After a vehicle is delivered to its owner, the cloud server continuously collects IMU static pose data for the first 5 minutes after the owner locks the car at frequently used parking locations (such as home garages and company parking lots). This data, combined with geographic location information, is used to cluster parking locations and gradually establish a "normal pose tolerance range" for each location. The normal pose tolerance range can be set as an interval ±0.5° above and below the average historical pose value of that location, and is dynamically adjusted as learning samples accumulate to adapt to parking scenarios with inherent tilt angles, such as ramp parking spaces.

[0076] Regarding the construction of the vehicle attitude reference model: the static reference attitude angle obtained from the factory static calibration and the normal attitude tolerance range learned from daily parking positions are used together to form the vehicle attitude reference model. When the system determines that the current parking position of the vehicle belongs to a learned location, the normal attitude tolerance range corresponding to that location is used as the anomaly detection reference first; when the parking position is an unlearned location, it reverts to the factory static reference attitude angle superimposed with the default tolerance range (e.g., ±1.0°) as the reference, ensuring that the system can work normally in any parking scenario.

[0077] S2: Receives remote vehicle locking command and activates anti-theft monitoring mode.

[0078] The core purpose of step S2 is to link the owner's remote locking operation with the activation of the anti-theft system, ensuring that the anti-theft monitoring starts immediately after the vehicle is physically locked, while controlling the system power consumption within the range required by the vehicle power management specifications through intermittent operation to ensure monitoring sensitivity.

[0079] Regarding the transmission chain of the vehicle locking command: The vehicle owner sends a remote locking command through an app on their mobile device. This command is uploaded to the cloud server via a 4G / 5G cellular network and then transmitted to the vehicle's remote communication terminal through a secure channel. After receiving the command, the remote communication terminal forwards the locking request to the body control module via the vehicle bus. The body control module then performs physical locking actions such as locking the doors and closing the windows to complete the vehicle locking process.

[0080] Regarding the broadcast mechanism of the anti-theft activation signal: After the body control module completes all locking actions and confirms that the vehicle is in a locked state, it broadcasts an "enter high-level anti-theft state" activation signal to the vehicle bus. The body domain controller continuously listens for bus messages on the vehicle bus. Upon receiving the aforementioned anti-theft activation signal, it immediately wakes up the six-axis inertial measurement unit and puts it into working mode, thus officially starting anti-theft monitoring.

[0081] Regarding the two phases of the intermittent operation mode: the anti-theft monitoring mode adopts an intermittent operation mode of "initial full-speed phase + periodic wake-up phase". Initial full-speed phase: the initial phase after anti-theft activation (e.g., the first 5 minutes), the six-axis inertial measurement unit runs continuously at full speed with the highest sampling frequency, and the anti-theft strategy module simultaneously performs dual-layer parallel anomaly detection to monitor the vehicle status with the highest sensitivity, preventing thieves from immediately committing theft when the vehicle is parked.

[0082] Periodic wake-up phase: If no abnormalities are detected during the initial full-speed phase, the system automatically switches to periodic wake-up mode. The six-axis inertial measurement unit (IMU) intermittently collects data according to a preset cycle (e.g., working at full speed for 5 seconds every 30 seconds). During the remaining time, the IMU enters a low-power standby state. Regarding power consumption control: Through the aforementioned intermittent operation, the increase in vehicle static current caused by the anti-theft system can be controlled within 5mA, and the vehicle static current is below 50mA, complying with vehicle power management specifications and preventing battery depletion due to prolonged vehicle parking.

[0083] S3: Acquire the vehicle's angular velocity and linear acceleration data, filter and fuse the angular velocity and linear acceleration data to obtain the vehicle's dynamic attitude angles and the processed triaxial linear acceleration.

[0084] The core purpose of step S3 is to process the raw angular velocity data and linear acceleration data output by the six-axis inertial measurement unit 1 through multi-level filtering and fusion to calculate high-precision dynamic attitude angles, angular velocities, and three-axis linear accelerations after gravity compensation, providing a reliable and low-noise data foundation for subsequent dual-layer parallel anomaly detection.

[0085] The six-axis inertial measurement unit (IMU) acquires angular velocity and linear acceleration data at a sampling frequency of no less than 50Hz, with a preferred sampling frequency of 100Hz. This ensures that the high-frequency micro-motion mode detection channel can capture tire disassembly characteristic vibration signals within the 5Hz to 15Hz frequency band, while simultaneously meeting the real-time monitoring requirements of the low-frequency attitude trajectory tracking channel for attitude angle change rate. The acquired raw data is transmitted in real-time to the sensor signal processing module of the vehicle domain controller via a dedicated serial peripheral interface or a high-speed controller area network (Controller Area Network) channel, with a transmission delay controllable to within 1ms.

[0086] The sensor signal processing module first performs low-pass filtering on the raw angular velocity and linear acceleration data to remove high-frequency interference components unrelated to theft, such as the sensor's own electrical noise and high-frequency structural vibrations of the vehicle body. The low-pass filter can be a second-order Butterworth low-pass digital filter with a cutoff frequency of 20Hz. This filter has a maximum flat amplitude-frequency response at the cutoff frequency, effectively suppressing high-frequency noise above 20Hz while maintaining good amplitude-frequency transmission characteristics in the 5Hz to 15Hz tire removal characteristic vibration frequency band, ensuring the accuracy of subsequent feature extraction.

[0087] The low-pass filtered data enters the Kalman filter algorithm module. This module uses the gravity vector measured by the accelerometer and the angle change obtained by the gyroscope integration as state variables to establish a dynamic state equation for the vehicle's attitude. Through a predictive-update iterative loop, it fuses the two types of sensor data in real time, outputting the following three types of fusion results: 1) dynamic attitude angles, including roll and pitch angles, reflecting the vehicle's current spatial attitude; 2) angular velocity, reflecting the rate of change of the vehicle's attitude; and 3) triaxial linear acceleration after gravity compensation, i.e., the actual vehicle acceleration obtained by subtracting the gravity component from the original accelerometer output, used for feature extraction in the high-frequency micro-motion mode detection channel. The attitude calculation accuracy can reach 0.1°, meeting the accuracy requirement of the low-frequency attitude trajectory tracking channel for a cumulative change detection threshold of 0.8°.

[0088] S4: Perform dual-layer parallel anomaly detection on dynamic attitude angles and triaxial linear acceleration to identify abnormal features of vehicle tire removal and vehicle lifting attitude deviation, and obtain anomaly detection results.

[0089] The design concept of step S4 is that tire removal behavior and vehicle lifting behavior exhibit distinct signal characteristics in the time and frequency domains. The former is manifested as a high-frequency periodic vibration envelope within a specific frequency band, while the latter is manifested as a low-speed unidirectional cumulative offset of the attitude angle. Therefore, two independent detection channels are used to specifically identify the above two types of characteristics. The two channels are executed in parallel without interfering with each other, and the detection results are output to the finite state machine for collaborative decision-making.

[0090] The dual-layer parallel anomaly detection includes high-frequency micro-motion pattern detection and low-frequency attitude trajectory tracking.

[0091] The high-frequency micro-motion mode detection involves: performing a frequency domain transformation on the triaxial linear acceleration to extract the envelope shape of the vibration energy on the time axis within a preset frequency band; matching the envelope shape with a pre-stored tire disassembly vibration template; and determining that an abnormal tire disassembly feature has been detected when the matching degree exceeds a preset threshold. The specific steps are as follows: First, frequency domain transformation. The frequency domain transformation of the gravity-compensated triaxial linear acceleration data can be performed using short-time Fourier transform (STFT) or wavelet analysis to convert the time-domain acceleration signal into a time-frequency domain representation, obtaining the two-dimensional time-frequency distribution of vibration energy across each frequency band. For STFT, a Hanning window can be used, with a window length of 256 sampling points and a frame shift of 64 sampling points; for wavelet analysis, the Daubechies wavelet basis can be used, with a decomposition level of up to 5 levels.

[0092] Second, envelope extraction. Extract the envelope shape of vibration energy on the time axis within a preset frequency band (which can be from 5Hz to 15Hz) from the time-frequency domain distribution. The envelope shape reflects the fluctuation and change law of vibration energy within the frequency band over time and is described as an energy-time curve.

[0093] Third, template matching. The extracted envelope shape is correlated with the tire removal vibration template pre-stored in the non-volatile memory of the vehicle body domain controller. The tire removal vibration template was obtained during the project development phase through laboratory bench simulation and real vehicle data acquisition. It describes the typical envelope shape (exhibiting 5 to 8 "bumps" with increasing amplitude) of the vibration energy in the 8Hz to 12Hz frequency band when the wheel nuts are loosened. The feature vector is described by wavelet coefficients or specific frequency band energy ratios. Correlation matching uses a normalized cross-correlation algorithm to calculate the similarity score between the measured envelope and the template.

[0094] Fourth, feature confirmation. When the similarity score exceeds a preset threshold, it is determined that an abnormal tire disassembly feature has been detected, and a tire disassembly abnormal feature flag is output to the finite state machine.

[0095] Low-frequency attitude trajectory tracking involves continuously monitoring changes in dynamic attitude angles at a preset update rate. When the dynamic attitude angle changes unidirectionally at a rate lower than a preset angular rate, and the cumulative change exceeds a preset angle value for a preset duration, a vehicle lifting attitude deviation characteristic is detected. The specific steps are as follows: First, update rate setting. The dynamic attitude angles (roll and pitch angles) of the Kalman filter fusion output are continuously sampled at a preset update rate (e.g., 1Hz) to establish an attitude angle time series for subsequent trend analysis.

[0096] Second, unidirectional change judgment. Real-time analysis of the attitude angle time series is performed to determine whether the attitude angle is continuously changing unidirectionally at a rate lower than a preset angular rate (e.g., 2° / s). The "continuous unidirectional change" requires that the direction of attitude angle change remains consistent across multiple consecutive sampling points (either all increasing or all decreasing), excluding random bidirectional fluctuations caused by road vibration or wind load.

[0097] Third, the cumulative change threshold judgment. Based on the confirmation that the attitude angle is continuously changing in one direction, the cumulative change of the attitude angle from the start of the change to the current time is calculated. When the cumulative change exceeds the preset angle value (e.g., 0.8°), the cumulative change threshold flag is triggered.

[0098] Fourth, duration threshold judgment. Based on the cumulative change exceeding the threshold, it is further judged whether the above unidirectional change process has lasted for a preset duration (e.g., 3 seconds) or more, in order to exclude short-term attitude deviation caused by brief disturbances of the people in the vehicle.

[0099] Fifth, feature confirmation. When the three conditions of unidirectional change judgment, cumulative change exceeding the threshold, and duration exceeding the threshold are simultaneously met, it is determined that a vehicle lifting posture deviation feature has been detected, and a vehicle lifting posture deviation feature flag is output to the finite state machine. The two detection channels execute in parallel within the anti-theft strategy module 22, each maintaining an independent data buffer and feature judgment logic, without interfering with each other. The feature output of one channel does not affect the normal operation of the other channel.

[0100] S5: Based on the anomaly detection results, determine the type of vehicle anomaly event by using a finite state machine to make a collaborative event decision.

[0101] The core purpose of step S5 is to perform time-series correlation analysis on the feature flags output by the dual-layer parallel anomaly detection channel. By requiring multiple anomaly features to corroborate each other in time, false triggers caused by single sporadic anomalies are eliminated, thereby accurately identifying compound theft events such as "illegal lifting" and "lifting and removing tires".

[0102] The anti-theft strategy module implements a five-state finite state machine. The five states are defined as follows: the idle state (IDLE) is the inactive state of the anti-theft system, and the six-axis inertial measurement unit is in a low-power standby mode and does not execute any detection logic. The monitoring status (MONITOR) indicates that the anti-theft system is activated and operating normally. The dual-layer parallel anomaly detection channel is continuously running, and the finite state machine continuously receives the characteristic flag outputs of the two channels. Suspected Lift State (SUSPECT_LIFT) is an intermediate state in which the system detects that the vehicle may be being lifted. In this state, the system maintains high sensitivity monitoring and starts timing. The suspected theft status (SUSPECT_THEFT) is a high-confidence abnormal state in which the system detects a combination of lifting and tire removal behavior; The alarm state (ALARM) is the final state in which the system confirms an abnormal event and generates an alarm command. The transition conditions between each state are as follows: After the system is powered on, the finite state machine enters the idle state; after receiving the anti-theft activation signal, the finite state machine switches from the idle state to the monitoring state and begins to continuously receive the dual-layer detection results.

[0103] In monitoring mode, when both the vehicle lifting posture deviation feature flag and the lifting start pulse feature flag are detected simultaneously, the system switches to a suspected lifting state and starts timing, with a timing duration of up to 10 seconds. The lifting start pulse feature is a single-peak pulse generated along the Z-axis at the moment the jack tray contacts the vehicle's bottom edge beam. This pulse duration can be 80ms to 120ms, and the amplitude can be 0.3g to 0.7g. The feature vector consists of the pulse width and peak value, obtained during the project development phase through laboratory bench simulation and real-vehicle data acquisition. It is pre-stored in the non-volatile memory of the vehicle domain controller 2 and identified in real time by the high-frequency micro-motion mode detection channel. In the suspected lifting state, if the same-side tire removal abnormality feature flag is detected, the system immediately switches to a suspected theft state without waiting for the timing to end. If no tire removal abnormality feature is detected after 10 seconds of timing, the system directly switches to the alarm state. Upon entering the suspected theft state, the finite state machine unconditionally switches to the alarm state and generates an alarm command.

[0104] Through the above state transition logic, the system requires that the vehicle lifting posture deviation characteristics, lifting start pulse characteristics, and tire removal abnormal characteristics corroborate each other in time sequence. Any single occasional abnormality (such as short-term posture fluctuation caused only by road vibration, or only a single impact pulse without continuous posture deviation) cannot trigger the state machine to enter the alarm state on its own, thereby effectively eliminating false triggering.

[0105] The timing during suspected lifting is implemented by a software timer within the anti-theft strategy module, with a timing accuracy of 100ms. During the timing period, a dual-layer parallel anomaly detection channel runs continuously, and a finite state machine continuously polls the detection results. If an abnormal feature of tire removal on the same side is detected during the timing process, the timing immediately stops and a state transition is triggered; if the timing expires, an alarm is triggered regardless of whether an abnormal feature of tire removal is detected, to prevent thieves from taking other destructive actions after lifting the vehicle without immediately removing the tires.

[0106] S6: When the abnormal vehicle event type meets the alarm conditions, generate an alarm data packet containing event type information and vehicle attitude information, and push a warning notification.

[0107] The core purpose of step S6 is to transform the alarm command generated by the finite state machine decision into an alarm data packet containing multi-dimensional and accurate information, and to push it securely and quickly to the car owner's mobile terminal through the complete vehicle-cloud-APP transmission link, so that the car owner can grasp the abnormal situation of the vehicle and take countermeasures as soon as possible.

[0108] When a vehicle abnormal event type meets the alarm conditions, the vehicle domain controller generates an alarm data packet, which specifically includes the following four types of information: First, there is event type information, which identifies the type of theft event that triggered the alarm, such as "suspected illegal lifting" or "suspected lifting and tire removal," represented by an enumeration code. Second, vehicle attitude information, including real-time roll and pitch angle values ​​at the time of alarm triggering, recorded with a resolution of 0.1°, intuitively reflecting the current tilt of the vehicle; Third, the timestamp of the event, using UTC time format, with an accuracy of milliseconds, is used to record the exact time of the alarm event, providing a time basis for subsequent accident identification; Fourth, the event's estimated location information is provided by a remote communication terminal based on the most recent valid global navigation satellite system positioning data or base station positioning data, marking the geographical coordinates of where the vehicle is parked, to help vehicle owners and the police quickly locate the vehicle.

[0109] After the alarm data packet is generated by the vehicle domain controller, it is sent to the remote communication terminal via the vehicle bus as a high-priority CAN message. The communication baud rate of the vehicle bus meets the real-time transmission requirements of vehicle signals, and the transmission delay can be controlled within 10ms. After receiving the alarm data packet, the remote communication terminal 4 uses an encrypted communication protocol (such as MQTT over TLS) to upload the encrypted data to the cloud server via the 4G / 5G cellular network. After receiving and decrypting the alarm data packet, the cloud server queries the owner account information bound to the vehicle VIN code and sends a high-priority push notification to the owner's mobile terminal through the push service to ensure that the warning information reaches the owner in a timely manner.

[0110] After receiving the warning notification, the vehicle owner's mobile terminal displays the vehicle's tilt status graphically. The current roll and pitch angle values ​​are presented intuitively through dynamic rendering of the vehicle's 3D model tilt angle. At the same time, detailed information such as event type, occurrence time, and estimated location of the event are displayed in text form.

[0111] The app interface also provides three quick operation entry points: one-click photo taking, horn honking and light flashing, and voice monitoring. When the car owner clicks "one-click photo taking," the in-vehicle camera can be remotely activated to take pictures of the vehicle's surroundings and send back the images; clicking "horn honking and light flashing" can control the vehicle's speakers and lights to issue warnings; clicking "voice monitoring" can use the in-vehicle microphone to listen to the ambient sounds around the vehicle in real time. All of the above quick operation commands are sent from the cloud server to the remote communication terminal for execution.

[0112] Example 3 This embodiment provides a vehicle, including a memory and a processor, as well as computer instructions stored in the memory and running on the processor. When the processor executes the computer instructions, it completes the remote anti-theft warning method based on a vehicle attitude sensor as described in Embodiment 1.

[0113] Example 4 This embodiment provides a computer-readable storage medium, characterized in that it is used to store computer instructions, which, when executed by a processor, complete the remote anti-theft warning method based on a vehicle attitude sensor as described in Embodiment 1.

[0114] Example 5 This embodiment provides a computer program product, which includes executable instructions stored in a computer-readable storage medium; wherein, when the vehicle's processor reads and executes the executable instructions from the computer-readable storage medium, it completes the remote anti-theft warning method based on a vehicle attitude sensor as described in Embodiment 1.

[0115] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0116] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0117] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment, whereby a series of operational steps are performed to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0118] The steps and methods involved in Embodiments 2 to 5 above correspond to those in Embodiment 1. For detailed implementation methods, please refer to the relevant description section of Embodiment 1. The descriptions of each embodiment in the above embodiments have different emphases. Parts not described in detail in a certain embodiment can be referred to the relevant descriptions of other embodiments.

[0119] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computer devices. Optionally, they can be implemented using computer-executable program code, thereby allowing them to be stored in a storage device for execution by a computer device, or they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. The present invention is not limited to any particular combination of hardware and software.

[0120] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A remote theft prevention warning method based on a vehicle attitude sensor, characterized by, Includes the following steps: Construct a vehicle attitude reference model, which is used to characterize the reference attitude range of the vehicle under normal parking conditions; Receives a remote vehicle locking command and activates the anti-theft monitoring mode; The vehicle's angular velocity and linear acceleration data are acquired, and the angular velocity and linear acceleration data are filtered and fused to obtain the vehicle's dynamic attitude angles and the processed triaxial linear acceleration. A two-layer parallel anomaly detection was performed on the dynamic attitude angle and triaxial linear acceleration to identify the abnormal features of vehicle tire removal and the vehicle lifting attitude deviation features, and the anomaly detection results were obtained. Based on the anomaly detection results, a finite state machine is used to make cooperative event judgments to determine the type of vehicle anomaly event; When a vehicle abnormal event type meets the alarm conditions, an alarm data packet containing event type information and vehicle attitude information is generated, and a warning notification is pushed.

2. The remote anti-theft early warning method based on a vehicle attitude sensor as described in claim 1, characterized in that, The steps for constructing a vehicle attitude baseline model are as follows: During the vehicle factory calibration phase, static reference attitude angles are obtained, including roll angle and pitch angle. By learning the posture data corresponding to the car owner's daily parking positions, a normal posture tolerance range corresponding to the parking position is established. The static reference attitude angle and the normal attitude tolerance range together constitute the vehicle attitude reference model.

3. The remote anti-theft early warning method based on a vehicle attitude sensor as described in claim 1, characterized in that, The angular velocity and linear acceleration data are filtered and fused as follows: Low-pass filtering is applied to angular velocity and linear acceleration data to suppress high-frequency noise, resulting in filtered data. The filtered data is fused using the Kalman filter algorithm to calculate the dynamic attitude angle, angular velocity, and triaxial linear acceleration after gravity compensation in real time.

4. The remote anti-theft early warning method based on a vehicle attitude sensor as described in claim 1, characterized in that, The dual-layer parallel anomaly detection includes high-frequency micro-motion mode detection and low-frequency attitude trajectory tracking; wherein, the high-frequency micro-motion mode detection is as follows: frequency domain transformation is performed on the three-axis linear acceleration to extract the envelope shape of the vibration energy in the preset frequency band on the time axis; the envelope shape is matched with the pre-stored tire disassembly vibration template, and when the matching degree exceeds the preset threshold, it is determined that the tire disassembly anomaly feature has been detected.

5. The remote anti-theft early warning method based on vehicle attitude sensor as described in claim 4, characterized in that, The low-frequency attitude trajectory tracking is as follows: continuously monitor the change of dynamic attitude angle at a preset update rate; when the dynamic attitude angle changes continuously in one direction at a rate lower than the preset angular rate, and the cumulative change exceeds the preset angle value and continues for a preset duration, it is determined that the vehicle lifting attitude deviation feature has been detected.

6. The remote anti-theft early warning method based on a vehicle attitude sensor as described in claim 1, characterized in that, The finite state machine includes an idle state, a monitoring state, a suspected lifting state, a suspected theft state, and an alarm state.

7. A remote anti-theft warning system based on a vehicle attitude sensor, characterized in that, Includes the following modules: The baseline establishment module is configured to: construct a vehicle attitude baseline model, which is used to characterize the baseline attitude range of the vehicle in a normal parking state; The anti-theft mode activation module is configured to: receive a remote vehicle locking command and activate the anti-theft monitoring mode; The data acquisition and fusion module is configured to: acquire the vehicle's angular velocity data and linear acceleration data, filter and fuse the angular velocity data and linear acceleration data to obtain the vehicle's dynamic attitude angle and the processed triaxial linear acceleration; The anomaly detection module is configured to perform dual-layer parallel anomaly detection on dynamic attitude angles and triaxial linear acceleration, respectively identifying abnormal features of vehicle tire removal and vehicle lifting attitude deviation, and obtaining anomaly detection results. The event decision module is configured to: determine the type of vehicle abnormal event by performing collaborative event decision through a finite state machine based on the anomaly detection results; The early warning push module is configured to generate an alarm data packet containing event type information and vehicle attitude information when the abnormal vehicle event type meets the alarm conditions, and push an early warning notification.

8. A vehicle, characterized in that, It includes a memory and a processor, as well as computer instructions stored in the memory and running on the processor, which, when executed by the processor, perform the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of claims 1-6.

9. A computer-readable storage medium, characterized in that, Used to store computer instructions, which, when executed by a processor, complete the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of claims 1-6.

10. A computer program product comprising executable instructions stored in a computer-readable storage medium; wherein, When the vehicle's processor reads executable instructions from a computer-readable storage medium and executes the executable instructions, it completes the remote anti-theft warning method based on a vehicle attitude sensor as described in any one of claims 1-6.

Images