A vehicle high-low order automatic driving self-adaptive switching strategy system and method

Abstract

The application discloses a kind of vehicle high-low order automatic driving self-adapting switching strategy system and method, including actuator performance dynamic evaluation module, scene risk grading evaluation module, driver state multidimensional monitoring module, switching decision module, in-vehicle multi-modal grading reminder module, outside collaborative grading reminder module, gradual control right handover module, monitoring and recording module, method, including normal degradation switching process, emergency degradation switching process, upgrade switching process, partial degradation switching process, driver refuses to take over processing flow;The safety of the application is significantly improved, the switching process is smoother, the safety factor inside and outside the vehicle is improved, the function availability is improved, the whole process is closed loop monitoring: from pre-switching to post-switching monitoring, a complete safety closed loop is formed, even in the case of driver takeover failure, the system can still provide auxiliary support.

Description

Technical Field

[0001] This invention relates to a system and method for adaptive switching strategy between high and low levels of autonomous driving in vehicles. Background Technology

[0002] With the official implementation of the "Administrative Measures for Access and Road Traffic of Intelligent Connected Vehicles," the process of granting access to Level 3 autonomous driving vehicles in my country has accelerated significantly. In the next 3-5 years, mass-produced vehicles will generally exhibit a technical form where Level 3 conditional autonomous driving coexists with Level 1 and Level 2 assisted driving functions.

[0003] The current technical solutions in the industry have the following core pain points and shortcomings:

[0004] 1. Single boundary between high and low order switching: The system only triggers switching based on environmental perception results, without considering the impact of actuator performance degradation on functional safety. This results in the L3 function being activated even when the actuator is in a sub-healthy state, posing a significant safety hazard.

[0005] 2. Lack of external warnings: Most switching logics only focus on warnings to the driver inside the vehicle, completely ignoring advance notifications to surrounding traffic participants such as vehicles, pedestrians, and non-motorized vehicles, which can easily lead to rear-end collisions, scrapes, and other accidents.

[0006] 3. Inadequate emergency response: When a driver refuses to take over or takes over fails, existing solutions often directly enter the minimum risk state (MRM), lacking tiered degrading and temporary maintenance mechanisms.

[0007] 4. Lack of post-switch evaluation: After the switch is completed, the quality of driver takeover is not continuously monitored, and timely intervention and assistance cannot be provided when the driver makes a mistake. Summary of the Invention

[0008] The purpose of this invention is to overcome the shortcomings of the prior art and provide a vehicle high-low level autonomous driving adaptive switching strategy system and method, which aims to solve the problem of safe, smooth and predictable switching between high-low level autonomous driving functions in vehicles with L3 and L3 and below driver assistance functions, while ensuring the safety of the driver in the vehicle and traffic participants outside the vehicle.

[0009] To achieve the above objectives, the present invention is implemented through the following technical solution:

[0010] A vehicle high-low level autonomous driving adaptive switching strategy system includes an actuator performance dynamic evaluation module, a scenario risk classification evaluation module, a driver state multi-dimensional monitoring module, a switching decision module, an in-vehicle multimodal classification reminder module, an external collaborative classification reminder module, a progressive control handover module, and a monitoring and recording module. The actuator performance dynamic evaluation module, scenario risk classification evaluation module, and driver state multi-dimensional monitoring module transmit data to the switching decision module. The driver state multi-dimensional monitoring module, in-vehicle multimodal classification reminder module, and external collaborative classification reminder module receive instructions issued by the switching decision module. The progressive control handover module and the monitoring and recording module all establish bidirectional communication connections with the switching decision module.

[0011] As a preferred embodiment, the actuator performance dynamic evaluation module is used to collect performance parameter data of each actuator, calculate the performance health index data of each actuator, and establish an actuator performance threshold matrix corresponding to L3 conditional autonomous driving, L2 assisted driving, and L1 assisted driving functions. When the actuator performance is lower than the current L3, L2, or L1 driving level requirements but higher than the next autonomous driving level requirements, the actuator performance dynamic evaluation module issues downgrade warning data and triggers smooth downgrade, smoothly transitioning from L3 to L2, L2 to L1, and L1 to manual driving, and outputs the smooth downgrade data to the switching decision module.

[0012] Preferably, the performance parameter data includes steering motor torque accuracy data, braking system response time data, and power system output delay data.

[0013] As a preferred option, the scenario risk classification and assessment module is used to collect environmental perception information data, construct environmental perception model data, and set scenario risk entropy value data and predicted risk level data, so as to provide a basis for scenario risk judgment for the level switching of L3 conditional autonomous driving, L2 assisted driving and L1 assisted driving.

[0014] As a preferred approach, the scenario risk entropy value is divided into four levels: low risk corresponds to a score range of 0-20, medium risk corresponds to a score range of 21-50, high risk corresponds to a score range of 51-80, and extremely high risk corresponds to a score range of 81-100.

[0015] As a preferred option, the driver status multi-dimensional monitoring module is used to collect driver facial expression data, gaze direction data, blink frequency data, and driver operation preparation status data. It also calculates driver takeover capability index data, verifies driver attention status data through a voice interaction system, and executes commands issued by the switching decision module to increase monitoring frequency.

[0016] As a preferred option, the in-vehicle multimodal hierarchical reminder module is used to execute hierarchical reminder commands issued by the switching decision module, and output visual, auditory, tactile, and olfactory reminder information corresponding to levels one to four in advance to the in-vehicle. If the level is level one, the reminder information is a text prompt on the central control screen combined with a gentle voice. If the level is level two, the reminder information is a flashing instrument panel, an increased voice volume, and a slight vibration of the seat. If the level is level three, the reminder information is a red warning on the HUD, a rapid voice, steering wheel vibration, and seat belt tightening. If the level is level four, the reminder information is flashing lights throughout the vehicle, a sharp alarm sound, strong seat vibration, and the release of a refreshing fragrance.

[0017] Preferably, the external collaborative hierarchical reminder module is used to execute the external reminder commands issued by the switching decision module. It outputs different prompts to different traffic participants outside the vehicle through the vehicle lights and external speakers. If the traffic participant is a vehicle behind, the prompt is the flashing frequency of the high-mounted brake light and the color change of the taillights. If the traffic participant is a vehicle in front, the prompt is the high beam headlights or dedicated switching lights. If the traffic participant is a pedestrian or non-motorized vehicle, the prompt is a voice prompt played by the external speaker combined with the flashing of the LED light strip on the side of the vehicle. The intensity of the prompt is positively correlated with the urgency of the switching and the risk level of the scenario.

[0018] As a preferred option, the progressive control handover module is used to execute the control handover instructions and control rollback instructions issued by the switching decision module, complete the transfer of control between L3 conditional autonomous driving, L2 assisted driving, L1 assisted driving and manual driving, and report the handover progress and driver operation status to the switching decision module.

[0019] The monitoring and recording module is used to execute the post-switching monitoring instructions issued by the switching decision module, collect and feed back driver operation quality data and operation deviation data to the switching decision module. When the driver's operation deviation is detected, the monitoring and recording module automatically activates the L2 level assisted driving function to correct the operation deviation, records the entire switching process data, and feeds back optimization data to the switching decision module to update the switching strategy.

[0020] The switching decision module receives and integrates data uploaded by various modules, comprehensively considers actuator performance health index, scene risk entropy value, and driver takeover capability index, and uses fuzzy inference algorithm to determine the switching timing, switching method, and switching level. It supports three switching modes: full switching, partial switching, and temporary maintenance mode between L3 conditional automated driving, L2 assisted driving, L1 assisted driving, and manual driving. The decision is made according to the priority rule that actuator safety takes precedence over scene safety, and scene safety takes precedence over driver status. After logical calculation, it generates control handover instructions, control reversal instructions, graded reminder instructions, external reminder instructions, and monitoring instructions, and sends them to the corresponding modules for execution, realizing adaptive and safe switching between high and low level conditions between L3 conditional automated driving, L2 assisted driving, L1 assisted driving, and manual driving.

[0021] A method for adaptive switching strategy between high and low levels of autonomous driving in vehicles includes a normal downgrade switching process, an emergency downgrade switching process, an upgrade switching process, a partial downgrade switching process, and a driver refusal to take over handling process.

[0022] The normal downgrade handover process is executed step-by-step, following the path from L3 conditional automated driving, L2 assisted driving, L1 assisted driving, to manual driving. It is triggered by the handover decision module based on actuator performance data collected by the actuator performance dynamic evaluation module and scene risk entropy data collected by the scene risk classification evaluation module. The triggering conditions are that the actuator performance is below the L3 threshold but above the L2 threshold, or the scene risk entropy value reaches a medium-risk level. This process consists of four stages, all coordinated and executed by the handover decision module: the pre-handover stage, the handover preparation stage, the gradual handover stage, and the handover completion stage.

[0023] The pre-switching phase involves the actuator performance dynamic evaluation module issuing degradation warning data and triggering smooth degradation, and feeding back the smooth degradation data to the switching decision module. The switching decision module controls the in-vehicle multimodal hierarchical reminder module to start the first-level reminder command, while the external collaborative hierarchical reminder module outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the system adjusts the vehicle status to a safe speed.

[0024] The switching preparation phase involves the switching decision module controlling the in-vehicle multimodal hierarchical reminder module to activate the secondary reminder information, while the external collaborative hierarchical reminder module outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the switching decision module sends a command to the driver status multidimensional monitoring module to increase the monitoring frequency.

[0025] The gradual handover phase involves the switching decision module controlling the in-vehicle multimodal hierarchical reminder module to activate three levels of reminder information. At the same time, the external collaborative hierarchical reminder module outputs different prompt information to different traffic participants outside the vehicle through vehicle lights and external speakers. Meanwhile, the gradual control handover module receives the control handover command issued by the switching decision module, and the control is linearly transferred at a rate of 20% / s.

[0026] The switching completion stage involves the switching decision module controlling the in-vehicle and out-of-vehicle reminders to stop and display the current driving level. At the same time, the monitoring and recording module executes the post-switching monitoring instructions issued by the switching decision module, continuously feeding back driver operation quality data and operation deviation data, and recording the entire switching process data.

[0027] The emergency downgrade switching process is initiated by the switching decision module based on the performance parameter data of each actuator collected by the actuator performance dynamic evaluation module, the performance health index data of each actuator calculated, and the environmental perception model data constructed by the scenario risk classification assessment module. The trigger condition is that the switching decision module determines that the actuator performance parameter data suddenly drops significantly or the environmental perception model data reaches an extremely high risk level. After triggering, the switching decision module sends a level 4 reminder message to the in-vehicle multimodal classification reminder module. At the same time, the external collaborative classification reminder module outputs different prompt messages to different traffic participants outside the vehicle through the vehicle lights and external speakers. The switching decision module immediately takes deceleration measures and then enters the rapid handover phase. The switching decision module executes the control handover command issued by the progressive control handover module, increasing the control transfer rate to 50% / s, achieving 50% control transfer in a short time. If the driver does not take over within the preset time, the switching decision module instructs the system to automatically switch to L2 level assisted driving mode. If the driver still does not take over, the switching decision module instructs the system to enter the minimum risk state MRM. The entire process is recorded by the monitoring and recording module, which records the switching and vehicle status data.

[0028] The upgrade and switching process is executed step by step according to the path from manual driving, L1-level assisted driving, L2-level assisted driving to L3-level conditional automated driving. The triggering conditions are the driver's active upgrade request, the actuator performance reaching the target level threshold, and the driver's attention status being verified as qualified through the voice interaction system. The switching decision module executes the graded reminder instructions issued by the in-vehicle multimodal graded reminder module, while the external collaborative graded reminder module executes the external reminder instructions issued by the switching decision module. Then, a gradual upgrade phase is carried out. The gradual control handover module executes the control handover instructions issued by the switching decision module, and the control is transferred linearly at a rate of 20% / s. After the upgrade is completed, the switching decision module controls the in-vehicle display of the current driving level and stops the external reminders. At the same time, the monitoring and recording module records the data of the entire upgrade process.

[0029] The partial downgrade switching process is triggered by the switching decision module based on the single actuator performance parameter data collected by the actuator performance dynamic evaluation module. The triggering condition is that the performance parameter data of a single actuator decreases, resulting in the limitation of some autonomous driving functions, while the performance of the remaining actuators meets the requirements of the current level. During the switch, the switching decision module instructs only the limited function dimension to be downgraded, while retaining the original autonomous driving level of the remaining function dimensions. At the same time, it instructs the in-vehicle multimodal hierarchical reminder module to issue a targeted takeover reminder instruction, and instructs the external collaborative hierarchical reminder module to output different prompt information to different traffic participants outside the vehicle through the vehicle lights and external speakers. When the actuator performance dynamic evaluation module detects that the performance of the faulty actuator has recovered to the corresponding level threshold, the switching decision module instructs the system to automatically restore the full-function operation of L3 conditional autonomous driving, and the monitoring and recording module records the function recovery data.

[0030] The driver refusal to take over handling process is coordinated and executed by the switching decision module. The switching decision module first instructs the in-vehicle multimodal hierarchical reminder module to issue a level three reminder command, while simultaneously verifying the driver's attention status data through the driver status multidimensional monitoring module. If no driver takeover operation is detected within 3 seconds, the switching decision module instructs the in-vehicle multimodal hierarchical reminder module to upgrade to a level four reminder command and controls the vehicle to slowly decelerate. If no driver takeover is detected within 5 seconds, the switching decision module instructs the progressive control handover module to automatically downgrade the system to L2 level assisted driving mode, temporarily maintaining vehicle operation for 30 seconds. If the driver status multidimensional monitoring module still does not detect driver takeover after 30 seconds, the switching decision module controls the vehicle to activate hazard lights and gradually decelerate to a stop, triggering an emergency rescue call. Throughout the process, the monitoring and recording module records data and simultaneously feeds back optimization data to the switching decision module to update the switching strategy.

[0031] The beneficial effects of this invention are as follows:

[0032] (1) Significantly improved safety: Through dynamic evaluation of actuator performance and scenario risk classification, potential risks can be predicted in advance, avoiding sudden loss of control due to actuator failure or environmental changes;

[0033] (2) Smoother switching process: The gradual control handover mechanism is adopted, which extends the control transfer time to 1-3 seconds, greatly reducing the driver's operational pressure and error rate;

[0034] (3) Improved safety inside and outside the vehicle: The vehicle-inside-outside coordinated graded reminder mechanism informs surrounding traffic participants in advance of changes in vehicle status, effectively reducing traffic accidents caused by switching.

[0035] (4) Improved functional availability: Supports partial degradation mode, which can retain some automatic driving functions when a certain actuator has a minor failure, thus improving the user experience;

[0036] (5) Full-process closed-loop monitoring: From pre-switch to post-switch monitoring, a complete safety closed loop is formed, and the system can still provide auxiliary support even in the event of driver takeover error. Attached Figure Description

[0037] Figure 1 This is a system block diagram of the present invention. Detailed Implementation

[0038] The technical solution of the present invention will be further described below with reference to the accompanying drawings:

[0039] like Figure 1 As shown, a vehicle high-low level autonomous driving adaptive switching strategy system includes an actuator performance dynamic evaluation module 100, a scenario risk classification evaluation module 200, a driver state multi-dimensional monitoring module 300, a switching decision module 400, an in-vehicle multimodal classification reminder module 500, an external collaborative classification reminder module 600, a progressive control handover module 700, and a monitoring and recording module 800. The actuator performance dynamic evaluation module 100, the scenario risk classification evaluation module 200, and the driver state multi-dimensional monitoring module 300 transmit data to the switching decision module 400. The driver state multi-dimensional monitoring module 300, the in-vehicle multimodal classification reminder module 500, and the external collaborative classification reminder module 600 receive instructions issued by the switching decision module 400. The progressive control handover module 700 and the monitoring and recording module 800 both establish bidirectional communication connections with the switching decision module 400.

[0040] like Figure 1 As shown, the actuator performance dynamic evaluation module 100 is used to collect performance parameter data of each actuator, calculate the performance health index data of each actuator, and establish an actuator performance threshold matrix corresponding to L3 level conditional autonomous driving, L2 level assisted driving, and L1 level assisted driving functions. When the actuator performance is lower than the current L3, L2, or L1 level driving level requirements but higher than the next autonomous driving level requirements, the actuator performance dynamic evaluation module 100 issues downgrade warning data and triggers smooth downgrade, smoothly transitioning from L3 level to L2 level, L2 level to L1 level, and L1 level to manual driving. The smooth downgrade data is output to the switching decision module 400. The performance parameter data includes steering motor torque accuracy data, braking system response time data, and power system output delay data.

[0041] like Figure 1As shown, the scenario risk classification and assessment module 200 is used to collect environmental perception information data, construct environmental perception model data, and set scenario risk entropy value data and predicted risk level data. It is used to provide a basis for scenario risk judgment for the level switching of L3 conditional autonomous driving, L2 assisted driving and L1 assisted driving. The scenario risk entropy value is divided into four levels: low risk corresponds to a score range of 0-20, medium risk corresponds to a score range of 21-50, high risk corresponds to a score range of 51-80, and extremely high risk corresponds to a score range of 81-100.

[0042] like Figure 1 As shown, the driver state multi-dimensional monitoring module 300 is used to collect driver facial expression data, gaze direction data, and blink frequency data through the DMS camera, and to collect driver operation preparation state data through the steering wheel grip force sensor and pedal sensor. At the same time, it calculates driver takeover ability index data (DTA), with a range of 0-100%. It also verifies driver attention state data through the voice interaction system and executes the command to increase the monitoring frequency issued by the switching decision module 400.

[0043] like Figure 1 As shown, the in-vehicle multimodal hierarchical reminder module 500 is used to execute the hierarchical reminder instructions issued by the switching decision module 400, and output visual, auditory, tactile and olfactory reminder information corresponding to the first to fourth levels to the in-vehicle. If the level is first level, the reminder information is a text prompt on the central control screen combined with a gentle voice. If the level is second level, the reminder information is a flashing instrument panel, an increased voice volume and a slight vibration of the seat. If the level is third level, the reminder information is a red warning on the HUD, a rapid voice, a steering wheel vibration and a tightening of the seat belt. If the level is fourth level, the reminder information is flashing lights throughout the vehicle, a sharp alarm sound, a strong vibration of the seat and the release of a refreshing fragrance.

[0044] like Figure 1 As shown, the external collaborative hierarchical reminder module 600 executes the external reminder commands issued by the switching decision module 400. It outputs different prompts to different traffic participants outside the vehicle through vehicle lights and external speakers. For example, external reminders are achieved based on the vehicle's LED light strips, headlights, brake lights, and external speakers. Differentiated reminder strategies are adopted for different traffic participants: if the traffic participant is a vehicle behind, the prompt is the flashing frequency of the high-mounted brake light and a change in taillight color; if the traffic participant is a vehicle in front, the prompt is the high / low beam headlights or a dedicated switching light; if the traffic participant is a pedestrian or non-motorized vehicle, the prompt is a voice prompt played through the external speaker combined with flashing LED light strips on the side of the vehicle. The intensity of the prompt is positively correlated with the urgency of the switching and the risk level of the scenario.

[0045] like Figure 1As shown, the progressive control handover module 700 is used to execute the control handover instructions and control rollback instructions issued by the switching decision module 400, complete the control transfer between L3 conditional automated driving, L2 assisted driving, L1 assisted driving and manual driving, and feed back the handover progress and driver operation status to the switching decision module 400.

[0046] like Figure 1 As shown, the monitoring and recording module 800 is used to execute the post-switching monitoring command issued by the switching decision module 400, collect and feed back driver operation quality data and operation deviation data to the switching decision module 400. When the driver's operation deviation is detected, the monitoring and recording module 800 automatically activates the L2 level assisted driving function to correct the operation deviation, records the entire process switching data, and feeds back optimization data to the switching decision module 400 to update the switching strategy.

[0047] like Figure 1 As shown, the switching decision module 400 receives and integrates data uploaded by various modules, comprehensively considers the actuator performance health index, scene risk entropy value, and driver takeover capability index, and uses a fuzzy inference algorithm to determine the switching timing, switching method, and switching level. It supports three switching modes: full switching, partial switching, and temporary maintenance mode between L3 conditional automated driving, L2 assisted driving, L1 assisted driving, and manual driving. The decision is made according to the priority rule that actuator safety takes precedence over scene safety, and scene safety takes precedence over driver status. After logical calculation, it generates control handover instructions, control rollback instructions, graded reminder instructions, external reminder instructions, and monitoring instructions, and sends them to the corresponding modules for execution, realizing adaptive and safe switching between high and low level conditions between L3 conditional automated driving, L2 assisted driving, L1 assisted driving, and manual driving.

[0048] A method for adaptive switching strategy between high and low levels of autonomous driving in vehicles includes a normal downgrade switching process, an emergency downgrade switching process, an upgrade switching process, a partial downgrade switching process, and a driver refusal to take over handling process.

[0049] The normal downgrade handover process is executed step-by-step according to the path from L3 conditional autonomous driving, L2 assisted driving, L1 assisted driving to manual driving. It is triggered by the handover decision module 400 based on actuator performance data collected by the actuator performance dynamic evaluation module 100 and scene risk entropy value data collected by the scene risk classification evaluation module 200. The triggering condition is that the actuator performance is below the L3 threshold but above the L2 threshold, or the scene risk entropy value reaches the medium risk level. This process is divided into four stages, all coordinated and executed by the handover decision module 400: the pre-handover stage, the handover preparation stage, the gradual handover stage, and the handover completion stage.

[0050] The pre-switching phase: 15-30 seconds in advance, the actuator performance dynamic evaluation module 100 issues a degradation warning data and triggers smooth degradation, and feeds back the smooth degradation data to the switching decision module 400. The switching decision module 400 controls the in-vehicle multimodal hierarchical reminder module 500 to start the first-level reminder command. At the same time, the external collaborative hierarchical reminder module 600 outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the system adjusts the vehicle status to a safe speed.

[0051] The switching preparation phase: 5-15 seconds in advance, the switching decision module 400 controls the in-vehicle multimodal hierarchical reminder module 500 to start the secondary reminder information. At the same time, the external collaborative hierarchical reminder module 600 outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the switching decision module 400 sends a command to the driver status multidimensional monitoring module 300 to increase the monitoring frequency.

[0052] The gradual handover phase involves the switching decision module 400 controlling the in-vehicle multimodal hierarchical reminder module 500 to activate three levels of reminder information. Simultaneously, the external collaborative hierarchical reminder module 600 outputs different prompts to different traffic participants outside the vehicle through vehicle lights and external speakers, such as rapidly flashing taillights and playing prompt sounds through external speakers. At the same time, the gradual control handover module 700 receives the control handover command issued by the switching decision module 400, and the control is linearly transferred at a rate of 20% / s.

[0053] The switching completion stage involves the switching decision module 400 controlling the in-vehicle and out-of-vehicle reminders to stop and display the current driving level. At the same time, the monitoring and recording module 800 executes the post-switching monitoring instructions issued by the switching decision module 400, continuously monitors for 30 seconds, continuously provides feedback on driver operation quality data and operation deviation data, and records the entire switching process data.

[0054] The emergency downgrade switching process is initiated by the switching decision module 400 based on the performance parameter data of each actuator collected by the actuator performance dynamic evaluation module 100, the performance health index data of each actuator calculated, and the environmental perception model data constructed by the scenario risk classification evaluation module 200. The trigger condition is that the switching decision module 400 determines that the actuator performance parameter data suddenly drops significantly or the environmental perception model data reaches an extremely high risk level. Upon triggering, the switching decision module 400 sends a level-four alert to the in-vehicle multimodal graded alert module 500. Simultaneously, the external collaborative graded alert module 600 outputs different prompts to different traffic participants outside the vehicle through vehicle lights and external speakers, such as all external lights flashing simultaneously and external... The loudspeaker emits a sharp alarm, and the switching decision module 400 immediately takes deceleration measures and activates the hazard lights. Then, it enters the rapid handover phase, where the switching decision module 400 executes the control handover command issued by the progressive control handover module 700, increasing the control transfer rate to 50% / s, achieving a 50% control transfer within 0.5s. If the driver does not take over within a preset time (e.g., 1s), the switching decision module 400 command system automatically switches to L2 level assisted driving mode. If the driver still does not take over after another time period (e.g., 3s), the switching decision module 400 command system enters the minimum risk state (MRM). The entire process is recorded by the monitoring and recording module 800, which records the switching and vehicle status data.

[0055] The upgrade and switching process is executed step by step according to the path from manual driving, L1-level assisted driving, L2-level assisted driving to L3-level conditional automated driving. The triggering conditions are the driver's active upgrade request, the actuator performance reaching the target level threshold, and the driver's attention status being verified as qualified through the voice interaction system. The switching decision module 400 executes the graded reminder instructions issued by the in-vehicle multimodal graded reminder module 500, and at the same time, the external collaborative graded reminder module 600 executes the external reminder instructions issued by the switching decision module 400. Then, a gradual upgrade phase is carried out. The gradual control handover module 700 executes the control handover instructions issued by the switching decision module 400, and the control is transferred linearly at a rate of 20% / s. After the upgrade is completed, the switching decision module 400 controls the in-vehicle display of the current driving level and stops the external reminders. At the same time, the monitoring and recording module 800 records the data of the entire upgrade process.

[0056] The partial downgrade switching process is triggered by the switching decision module 400 based on the single actuator performance parameter data collected by the actuator performance dynamic evaluation module 100. The triggering condition is that the performance parameter data of a single actuator decreases, resulting in the limitation of some autonomous driving functions, while the performance of the remaining actuators meets the requirements of the current level. During the switch, the switching decision module 400 instructs only the limited function dimension to be downgraded, while retaining the original autonomous driving level of the remaining function dimensions. At the same time, it instructs the in-vehicle multimodal hierarchical reminder module 500 to issue targeted takeover reminder instructions, such as an in-vehicle reminder clearly informing the driver "Please take over the steering wheel, the system will maintain the vehicle speed and distance," and instructs the external collaborative hierarchical reminder module 600 to output different prompt information to different traffic participants outside the vehicle through the vehicle lights and external speakers. When the actuator performance dynamic evaluation module 100 detects that the performance of the faulty actuator has recovered to the corresponding level threshold, the switching decision module 400 instructs the system to automatically restore the full-function operation of L3 level conditional autonomous driving, and the monitoring and recording module 800 records the function recovery data.

[0057] The driver refusal to take over handling process is coordinated and executed by the switching decision module 400. The switching decision module 400 first instructs the in-vehicle multimodal hierarchical reminder module 500 to issue a level 3 reminder command, and at the same time verifies the driver's attention status data through the driver status multidimensional monitoring module 300. If no driver takeover operation is detected within 3 seconds, the switching decision module 400 instructs the in-vehicle multimodal hierarchical reminder module 500 to upgrade to a level 4 reminder command and controls the vehicle to slow down slowly. If no driver takeover is detected within 5 seconds, the switching decision module 400 instructs the progressive control handover module 700 to automatically downgrade the system to L2 level assisted driving mode and temporarily maintain vehicle operation for 30 seconds. If the driver status multidimensional monitoring module 300 still does not detect driver takeover after 30 seconds, the switching decision module 400 controls the vehicle to turn on the hazard lights and gradually slow down to a stop, and triggers the emergency rescue call operation. Throughout the process, the monitoring and recording module 800 records the data and synchronously feeds back optimization data to the switching decision module 400 to update the switching strategy.

[0058] This invention possesses significant technical advantages by constructing a high-low level autonomous driving adaptive switching system that integrates "multi-module collaboration, full-scenario coverage, and progressive handover": First, the modules have clear division of labor and deep collaboration. Three input modules—dynamic evaluation of actuator performance, scenario risk classification, and multi-dimensional monitoring of driver status—provide comprehensive data support. The switching decision module achieves accurate decision-making through fuzzy inference algorithms. Combined with in-vehicle and out-of-vehicle graded reminders, progressive control handover, and monitoring and recording modules, a complete closed loop of "perception-decision-execution-feedback" is formed, ensuring rigorous switching logic. Second, the switching modes are flexibly adaptable to all scenarios, covering five processes: normal downgrade, emergency downgrade, upgrade, partial downgrade, and driver refusal to take over. It supports smooth transitions from L3 to L2 to L1 to manual driving and can also respond quickly to emergency scenarios such as sudden malfunctions and extremely high risks. The partial downgrade mode further achieves "partial functional adjustment..." The system features refined control that allows for "no need to downgrade the entire level," balancing safety and driving experience. Thirdly, it boasts ample safety redundancy, with in-vehicle four-level multimodal alerts combined with differentiated external collaborative prompts. This addresses the pain points of insufficient driver awareness and information asymmetry among traffic participants during the transition process. The progressive control transfer rate can be dynamically adjusted, and combined with driver deviation correction, minimum risk state (MRM), and emergency rescue mechanisms, it minimizes transition risks. Fourthly, it combines intelligence and practicality. Designs such as Driver Takeover Ability Index (DTA) calculation, dynamic monitoring frequency enhancement, and iterative updates to transition strategies enable real-time adaptation to driver status, actuator performance, and scenario risks. The differentiated alert strategies adapted to different traffic participants further enhance the feasibility of the solution, effectively filling the gaps in existing autonomous driving transition technologies, which suffer from "single scenario adaptation, insufficient safety alerts, and abrupt control transfers."

[0059] The actuator performance requirements and threshold design references are shown in the table below. The values ​​can be adjusted adaptively according to the actual situation.

[0060]

[0061] Note: When the actuator performance parameters are between two levels, the system automatically downgrades to the lower level; when the actuator performance is below the L1 level requirement, the system immediately alerts the driver to take over and prepares to enter the minimum risk state.

[0062] Practical usage examples:

[0063] The vehicle was traveling at 120 km / h on the highway, and the Level 3 autonomous driving function was activated normally. During the journey, the actuator performance monitoring module detected that the steering motor torque accuracy decreased from ±3% to ±8%, which is lower than the Level 3 requirement (≤±5%) but higher than the Level 2 requirement (≤±10%).

[0064] The system triggers the normal degradation process, as described below:

[0065] (1) Activate the in-vehicle Level 1 reminder 20 seconds in advance: The central control screen displays "You are about to be downgraded to Level 2 assisted driving. Please prepare to take over the steering wheel" and plays a gentle voice message.

[0066] (2) Send pre-downgrade information to vehicles within a 100-meter range ahead and behind via the vehicle external collaborative graded reminder module 600;

[0067] (3) Activate the in-vehicle secondary reminder 10 seconds in advance: the yellow indicator light on the instrument panel flashes, the voice volume increases, and the seat vibrates slightly.

[0068] (4) The taillights on the outside of the vehicle begin to flash slowly at a frequency of 1Hz;

[0069] (5) Activate the three-level in-vehicle reminder 2 seconds in advance: HUD displays a red warning, the steering wheel vibrates, and the seat belt tightens;

[0070] (6) The exterior taillights flash rapidly at a frequency of 3Hz, and the external speaker plays a warning sound that “the vehicle is about to be downgraded, please keep a safe distance”;

[0071] (7) The system control is transferred linearly at a rate of 20% / s, and the driver gains 100% control after 5s;

[0072] (8) After the switch is completed, the vehicle displays "Currently in L2 level assisted driving mode";

[0073] (9) The monitoring and recording module 800 starts and continuously monitors the driver's operation for 30 seconds. It exits after no abnormality is found.

[0074] Compared with existing technologies, this patent proposes a comprehensive strategy for switching between high and low levels of autonomous driving, encompassing "dynamic actuator performance evaluation, scenario risk grading, driver setup monitoring, in-vehicle and out-of-vehicle collaborative alerts, gradual control handover, and closed-loop monitoring after the switch." This maximizes safety during the transition from L3 autonomous driving to below L3 assisted driving, avoiding ambiguity in traffic liability caused by the switch. It also enhances the safety of passengers and other road users in high-level autonomous driving scenarios.

[0075] It should be noted that the above examples are merely one specific embodiment of the present invention. Obviously, the present invention is not limited to the above embodiments, and many variations are possible. In short, all variations that can be directly derived or conceived by those skilled in the art from the content disclosed in this invention should be considered within the scope of protection of this invention.

Claims

1. A vehicle high- and low-level autonomous driving adaptive switching strategy system, characterized in that: The system includes an actuator performance dynamic evaluation module (100), a scenario risk classification evaluation module (200), a driver status multi-dimensional monitoring module (300), a switching decision module (400), an in-vehicle multimodal classification reminder module (500), an external collaborative classification reminder module (600), a progressive control handover module (700), and a monitoring and recording module (800). The actuator performance dynamic evaluation module (100), the scenario risk classification evaluation module (200), and the driver status multi-dimensional monitoring module (300) transmit data to the switching decision module (400). The driver status multi-dimensional monitoring module (300), the in-vehicle multimodal classification reminder module (500), and the external collaborative classification reminder module (600) receive instructions issued by the switching decision module (400). The progressive control handover module (700) and the monitoring and recording module (800) all establish bidirectional communication connections with the switching decision module (400).

2. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 1, characterized in that: The actuator performance dynamic evaluation module (100) is used to collect the performance parameter data of each actuator, calculate the performance health index data of each actuator, and establish the actuator performance threshold matrix corresponding to L3 level conditional autonomous driving, L2 level assisted driving, and L1 level assisted driving functions. When the actuator performance is lower than the current L3 level, L2 level or L1 level driving level requirement but higher than the next autonomous driving level requirement, the actuator performance dynamic evaluation module (100) issues downgrade warning data and triggers smooth downgrade, smoothly transitioning from L3 level to L2 level, L2 level to L1 level, and L1 level to manual driving, and outputs the smooth downgrade data to the switching decision module (400).

3. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 2, characterized in that: The performance parameter data includes steering motor torque accuracy data, braking system response time data, and power system output delay data.

4. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 3, characterized in that: The scenario risk classification and assessment module (200) is used to collect environmental perception information data, construct environmental perception model data, and set scenario risk entropy value data and predicted risk level data, so as to provide a basis for scenario risk judgment for the level switching of L3 conditional autonomous driving, L2 assisted driving and L1 assisted driving.

5. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 4, characterized in that: The scenario risk entropy value is divided into four levels: low risk corresponds to a score range of 0-20, medium risk corresponds to a score range of 21-50, high risk corresponds to a score range of 51-80, and extremely high risk corresponds to a score range of 81-100.

6. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 5, characterized in that: The driver status multi-dimensional monitoring module (300) is used to collect driver facial expression data, gaze direction data, blink frequency data, and driver operation preparation status data. It also calculates driver takeover capability index data and verifies driver attention status data through a voice interaction system. Furthermore, it is used to execute the command to increase monitoring frequency issued by the switching decision module (400).

7. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 6, characterized in that: The in-vehicle multimodal hierarchical reminder module (500) is used to execute the hierarchical reminder instructions issued by the switching decision module (400), and output visual, auditory, tactile and olfactory reminder information corresponding to the first to fourth levels to the in-vehicle in advance. If the level is first level, the reminder information is a text prompt on the central control screen combined with a gentle voice. If the level is second level, the reminder information is a flashing instrument panel, an increased voice volume and a slight vibration of the seat. If the level is third level, the reminder information is a red warning on the HUD, a rapid voice, a vibration of the steering wheel and a tightening of the seat belt. If the level is fourth level, the reminder information is flashing lights throughout the vehicle, a sharp alarm sound, a strong vibration of the seat and the release of a refreshing fragrance.

8. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 7, characterized in that: The external collaborative hierarchical reminder module (600) is used to execute the external reminder command issued by the switching decision module (400). It outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. If the traffic participant is a vehicle behind, the prompt information is the flashing frequency of the high-mounted brake light and the color change of the taillight. If the traffic participant is a vehicle in front, the prompt information is the high beam headlight or the dedicated switching light. If the traffic participant is a pedestrian or non-motorized vehicle, the prompt information is the voice prompt played by the external speaker combined with the flashing of the LED light strip on the side of the vehicle body. The intensity of the prompt information is positively correlated with the urgency of the switching and the risk level of the scenario.

9. The adaptive switching strategy system for high- and low-level autonomous driving of a vehicle according to claim 8, characterized in that: The progressive control handover module (700) is used to execute the control handover instructions and control rollback instructions issued by the switching decision module (400), complete the control transfer between L3 conditional automated driving, L2 assisted driving, L1 assisted driving and manual driving, and report the handover progress and driver operation status to the switching decision module (400). The monitoring and recording module (800) is used to execute the post-switching monitoring command issued by the switching decision module (400), collect and feed back driver operation quality data and operation deviation data to the switching decision module (400). When the driver's operation deviation is detected, the monitoring and recording module (800) automatically activates the L2 level assisted driving function to correct the operation deviation, records the switching data of the whole process, and feeds back the optimization data to the switching decision module (400) to update the switching strategy. The switching decision module (400) is used to receive and integrate the data uploaded by each module, comprehensively consider the actuator performance health index, scene risk entropy value and driver takeover capability index, and use fuzzy inference algorithm to determine the switching timing, switching method and switching level. It supports three switching modes: full switching, partial switching and temporary maintenance mode between L3 conditional autonomous driving, L2 assisted driving, L1 assisted driving and manual driving. It makes decisions according to the priority rule that actuator safety takes precedence over scene safety and scene safety takes precedence over driver status. After logical calculation, it generates control handover instructions, control rollback instructions, graded reminder instructions, vehicle external reminder instructions and monitoring instructions, and sends them to the corresponding modules for execution to realize adaptive safe switching between high and low level conditions between L3 conditional autonomous driving, L2 assisted driving, L1 assisted driving and manual driving.

10. A method for adaptive switching strategy between high and low levels of autonomous driving in a vehicle, based on the adaptive switching strategy system for high and low levels of autonomous driving as described in any one of claims 1 to 9, characterized in that, This includes normal downgrade handover procedures, emergency downgrade handover procedures, upgrade handover procedures, partial downgrade handover procedures, and procedures for handling driver refusal to take over. The normal downgrade switching process is executed step by step according to the path from L3 conditional autonomous driving, L2 assisted driving, L1 assisted driving to manual driving. It is triggered by the switching decision module (400) based on the actuator performance data collected by the actuator performance dynamic evaluation module (100) and the scene risk entropy value data collected by the scene risk classification evaluation module (200). The triggering condition is that the actuator performance is lower than the L3 threshold and higher than the L2 threshold, or the scene risk entropy value reaches the medium risk level. The process is divided into four stages, all of which are coordinated and executed by the handover decision module (400) issuing instructions. The four stages are the pre-handover stage, the handover preparation stage, the gradual handover stage, and the handover completion stage. The pre-switching phase is that the actuator performance dynamic evaluation module (100) issues downgrade warning data and triggers smooth downgrade, and feeds back the smooth downgrade data to the switching decision module (400). The switching decision module (400) controls the in-vehicle multimodal hierarchical reminder module (500) to start the first-level reminder command. At the same time, the external collaborative hierarchical reminder module (600) outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the system adjusts the vehicle status to a safe speed. The switching preparation phase involves the switching decision module (400) controlling the in-vehicle multimodal hierarchical reminder module (500) to activate the secondary reminder information, while the external collaborative hierarchical reminder module (600) outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. At the same time, the switching decision module (400) sends an instruction to the driver status multidimensional monitoring module (300) to increase the monitoring frequency. The gradual handover phase involves the switching decision module (400) controlling the in-vehicle multimodal hierarchical reminder module (500) to activate the three-level reminder information. At the same time, the external collaborative hierarchical reminder module (600) outputs different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. Meanwhile, the gradual control handover module (700) receives the control handover command issued by the switching decision module (400), and the control is linearly transferred at a rate of 20% / s. The switching completion stage is that the switching decision module (400) controls the in-vehicle and out-of-vehicle reminders to stop and display the current driving level. At the same time, the monitoring and recording module (800) executes the post-switching monitoring instructions issued by the switching decision module (400), continuously feeds back the driver's operation quality data and operation deviation data, and records the entire switching process data. The emergency downgrade switching process is initiated by the switching decision module (400) based on the performance parameter data of each actuator collected by the actuator performance dynamic evaluation module (100), the performance health index data of each actuator calculated, and the environmental perception model data constructed by the scenario risk classification evaluation module (200). The trigger condition is that the switching decision module (400) judges that the actuator performance parameter data suddenly drops sharply or the environmental perception model data reaches an extremely high risk level. After triggering, the switching decision module (400) sends a level 4 reminder message to the in-vehicle multimodal classification reminder module (500). At the same time, the out-of-vehicle collaborative classification reminder module (600) outputs different prompt messages to different traffic participants outside the vehicle through the vehicle body lights and external speakers. The switching decision module (400) immediately takes deceleration measures and then enters the rapid handover stage. The switching decision module (400) executes the control handover instruction issued by the progressive control handover module (700) to increase the control transfer rate to 50% / s, so as to complete 50% control transfer in a short time. If the driver fails to take over within the preset time, the switching decision module (400) command system will automatically switch to L2 level assisted driving mode. If the driver still fails to take over, the switching decision module (400) command system will enter the minimum risk state MRM. The switching and vehicle status data will be recorded by the monitoring and recording module (800) throughout the process. The upgrade and switching process is executed step by step according to the path from manual driving, L1-level assisted driving, L2-level assisted driving to L3-level conditional automated driving. The triggering conditions are the driver's active upgrade request, the actuator performance reaching the target level threshold, and the driver's attention status being verified as qualified through the voice interaction system. The switching decision module (400) executes the graded reminder instruction issued by the in-vehicle multimodal graded reminder module (500), and at the same time, the external collaborative graded reminder module (600) executes the external reminder instruction issued by the switching decision module (400). Then, a gradual upgrade phase is carried out. The gradual control handover module (700) executes the control handover instruction issued by the switching decision module (400). The control is transferred linearly at a rate of 20% / s. After the upgrade is completed, the switching decision module (400) controls the in-vehicle display of the current driving level and the external reminder to stop. At the same time, the monitoring and recording module (800) records the data of the entire upgrade process. The partial downgrade switching process is triggered by the switching decision module (400) based on the single actuator performance parameter data collected by the actuator performance dynamic evaluation module (100). The triggering condition is that the single actuator performance parameter data decreases, resulting in the limitation of some autonomous driving functions, while the performance of the remaining actuators meets the requirements of the current level. During the switching, the switching decision module (400) instructs only the limited function dimension to be downgraded, while retaining the original autonomous driving level of the remaining function dimensions. At the same time, it instructs the in-vehicle multimodal hierarchical reminder module (500) to issue a targeted takeover reminder instruction, and instructs the external collaborative hierarchical reminder module (600) to output different prompt information to different traffic participants outside the vehicle through the vehicle body lights and external speakers. When the actuator performance dynamic evaluation module (100) detects that the performance of the faulty actuator has recovered to the corresponding level threshold, the switching decision module (400) instructs the system to automatically restore the full-function operation of L3 level conditional autonomous driving, and the monitoring and recording module (800) records the function recovery data. The driver refusal to take over handling process is coordinated and executed by the switching decision module (400). The switching decision module (400) first instructs the in-vehicle multimodal hierarchical reminder module (500) to issue a level 3 reminder instruction, and at the same time verifies the driver's attention status data through the driver status multidimensional monitoring module (300). If no driver takeover operation is detected within 3 seconds, the switching decision module (400) instructs the in-vehicle multimodal hierarchical reminder module (500) to upgrade to a level 4 reminder instruction, and controls the vehicle to slow down slowly. If no driver takeover operation is detected within 5 seconds... The switching decision module (400) instructs the progressive control handover module (700) to automatically downgrade the system to L2 level assisted driving mode and temporarily maintain vehicle operation for 30 seconds. If the multi-dimensional driver status monitoring module (300) still does not detect driver takeover after 30 seconds, the switching decision module (400) controls the vehicle to turn on hazard lights and gradually decelerate to a stop, and triggers the emergency rescue call operation. The monitoring and recording module (800) records the data throughout the process and synchronously feeds back the optimization data to the switching decision module (400) to update the switching strategy.

Images