Abnormality detection and degradation stability control method for two-wheeled vehicle

Abstract

The application discloses an abnormality detection and degradation stability control method of a two-wheeled vehicle, belongs to the technical field of electronic control and safety of two-wheeled vehicles, and comprises the steps of synchronously collecting sensor monitoring data, actuator operation data and vehicle-mounted network communication data of the two-wheeled vehicle, realizing full-dimension detection and accurate positioning of three types of abnormalities of sensors, communication links and actuators without relying on mechanical gyroscopic torque stabilizing devices, combining abnormality grades and types to match a graded and scenario-based degradation strategy, constructing a lateral and longitudinal collaborative stability control model, designing an actuator abnormality takeover compensation strategy and an emergency stability degradation strategy, simultaneously adjusting control parameters dynamically through adaptive acquisition of feedback data, realizing full-process closed-loop optimization of abnormality detection-degradation control-stability maintenance, effectively dealing with single and combined abnormal working conditions, and significantly improving controllability and driving safety of the vehicle under abnormal working conditions.

Description

Technical Field

[0001] This invention belongs to the field of electronic control and safety technology for two-wheeled vehicles, specifically referring to a method for anomaly detection and degraded stability control of two-wheeled vehicles. Background Technology

[0002] As two-wheeled vehicles rapidly develop towards electrification, intelligence, and connectivity, their core control systems are becoming increasingly complex. The communication links between sensors, actuators, and various control units have become crucial to ensuring the normal operation of the vehicles. However, in actual driving, two-wheeled vehicles often face various abnormal conditions such as sensor drift and failure, actuator jamming and response delay, and communication link interference and interruption. If these abnormalities cannot be detected and effectively addressed in a timely manner, they can easily lead to vehicle instability and sudden changes in driving status, seriously threatening driving safety. However, existing methods for anomaly detection and degraded stability control of two-wheeled vehicles still have certain shortcomings. Current technologies lack systematicity and closed-loop management; data acquisition is fragmented, from a single source, and asynchronous in time; and crude preprocessing methods lead to inconsistent data quality. Anomaly identification methods are simplistic, with poor adaptability to various types of anomalies from sensors, communications, and actuators, easily resulting in missed or false positives. Furthermore, anomaly localization is vague, and grading is unreasonable, making it difficult to match targeted degraded strategies. Degraded modes are singular and insufficiently adaptable to anomaly types and levels, and control authority allocation is unreasonable. Degraded control models are also highly homogenized, hindering effective implementation. The vehicle exhibits poor adaptability to various abnormal scenarios, lacks reliable redundancy guarantees and verification mechanisms for command transmission, and is prone to transmission failures. Furthermore, the absence of an effective execution feedback loop makes it impossible to verify and correct control effects in real time, hindering the effective suppression of vehicle roll, yaw, and instability during rapid acceleration and deceleration. Degraded control lacks comprehensive state monitoring methods, and the fixed control parameters cannot adapt to dynamic changes in abnormal states, resulting in insufficient stability and overall inability to guarantee the safety and stability of the vehicle under abnormal operating conditions. Therefore, this paper proposes an anomaly detection and degraded stability control method for two-wheeled vehicles. Summary of the Invention

[0003] The purpose of this invention is to provide an anomaly detection and degraded stability control method for two-wheeled vehicles to solve the problems mentioned in the background art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a method for anomaly detection and degraded stability control of a two-wheeled vehicle, comprising the following steps: S1. Synchronously collect sensor monitoring data, actuator operation data and communication data between various control units of the two-wheeled vehicle to complete the initial data collection; S2. Perform noise reduction, completion, and standardization preprocessing on the collected data to generate a qualified set of data to be tested; S3. Based on the preprocessed data, perform multi-dimensional anomaly identification on sensors, communication links, and actuators, and mark suspected anomalies and their preliminary types. S4. Based on the suspected abnormal results, accurately locate the source of the abnormality and classify it according to the degree of impact of the abnormality on driving safety. S5. Based on the anomaly classification results, match the preset degradation strategy library to determine the target degradation mode and control permission allocation scheme; S6. Based on the target degradation mode, construct a vehicle stability control model, output adapted degradation control commands, and drive actuator actions. S7. Real-time acquisition of vehicle status feedback data after downgrade control, and dynamic adjustment of control parameters; It is achieved without relying on mechanical gyroscope torque stabilization devices.

[0005] Preferably, the sensor monitoring data includes at least vehicle attitude parameters and motion state parameters; The actuator operating data includes at least the actuator output value and response delay; The vehicle network communication data includes message data and / or communication status data between the controller, sensors and actuators; And timestamp calibration is used to achieve synchronous aggregation of data from different sources.

[0006] Preferably, the preprocessing includes denoising using at least one of sliding window filtering, low-pass filtering, Kalman filtering, or median filtering; Missing values ​​are filled using at least one of interpolation or historical estimation. Dimensional unification is achieved using at least one of normalization or standardization methods. The standardization formula is as follows: , In the formula, For standardized data, This is the original data. The mean of the original data sequence. denoted as the standard deviation of the original data sequence.

[0007] Preferably, the anomaly detection includes: Perform threshold verification, trend analysis, rate of change detection, and / or consistency detection on sensor data to identify sensor anomalies; Perform frame loss detection, timeout detection, CRC check failure detection, and / or communication delay detection on the vehicle network communication data to determine communication link anomalies; The actuator operation data is processed by executing instructions—response residual detection and / or response timing detection—to determine actuator anomalies, implemented as follows: , In the formula, For the amount of data change, For time-varying quantities, if the data exceeds the threshold range or the rate of change shows an unexpected abrupt change, it is marked as a suspected sensor anomaly; communication anomaly identification includes signal strength detection, transmission delay detection, and data integrity detection. It verifies whether the transmitted data is lost through a checksum. If the communication signal strength is lower than a preset threshold, the transmission delay is higher than a preset threshold, or there is a data integrity problem, it is marked as a suspected communication anomaly; actuator anomaly identification compares the deviation between the actual output parameters and the theoretical output parameters of the actuator. If the deviation continues to exceed the preset allowable range, it is marked as a suspected actuator anomaly; all suspected anomalies are marked with a preliminary anomaly type.

[0008] Preferably, the anomaly source localization is achieved based on data tracing and / or fault tree analysis; The anomaly classification is determined at least based on the degree of impact of the anomaly on vehicle stability, its controllability and / or recoverability, and severe anomalies correspond to anomalies that may lead to vehicle instability.

[0009] Preferably, the degradation strategy library includes multiple degradation modes, with different degradation modes corresponding to different anomaly types and anomaly levels; Among them, minor anomalies correspond to monitoring mode; moderate anomalies correspond to speed limit and amplitude limit mode; and severe anomalies correspond to safe speed reduction or parking stability mode. Furthermore, in the event of severe anomalies, the control permission allocation scheme forcibly sets the control permissions related to attitude stabilization control to the highest priority and restricts the control permissions of non-core functions.

[0010] Preferably, the degraded stability control command is generated by constructing a collaborative control model of lateral stability control and longitudinal stability control; Lateral stability control includes adjusting steering control output and / or auxiliary support control output to suppress roll tendency; longitudinal stability control includes limiting drive control output and / or applying braking control output to achieve smooth deceleration. The collaborative control model employs at least one of proportional-integral-derivative (PID) control, model predictive control (MPC), or robust control. Furthermore, when an abnormality is detected in the communication link, the stability control command is switched to the redundant communication link for transmission. The network message of the stability control command has a higher priority than the drive and steering control messages, in order to preempt the vehicle network communication resources. The control model uses the PID control algorithm to calculate the core control commands, and the control formula is as follows: , In the formula, To control the output value of the command, This refers to the deviation between the actual state of the vehicle and the target stable state. This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients; After the control command is generated, it is first transmitted through a communication link that has been confirmed to be free of abnormalities. If a communication abnormality is found, it is immediately switched to a redundant communication link, and a CRC check code is added to ensure the integrity of the command transmission. After the actuator receives the command, it needs to provide feedback on the execution status within a preset response time. If the actuator is detected to have not acted according to the command, the control model will adjust the command parameters in real time.

[0011] Preferably, the status feedback data includes at least attitude parameters, vehicle speed parameters, actuator response parameters, and communication status parameters; The sampling frequency is adaptively adjusted according to the anomaly level; Based on the feedback data, deviation trend assessment and / or cumulative error assessment are performed. When the deviation exceeds the threshold, the degradation control parameters are updated. The collected feedback data is compared with the preset stability threshold range in the corresponding degradation mode to calculate the cumulative deviation between the actual state and the target stable state. ; If the cumulative deviation is within the allowable range, maintain the current downgraded control parameters; if the cumulative deviation exceeds the allowable range, make targeted adjustments based on the marked anomaly type: for sensing deviations caused by sensor anomalies, optimize the weights of the state estimation model to weaken the impact of abnormal sensor data.

[0012] Preferably, the anomaly detection further includes a cross-dimensional correlation verification mechanism to identify suspected anomalies by verifying the logical consistency between sensor data, communication data, and actuator data; When a suspected sensor malfunction is detected, redundant sensors are invoked for cross-validation to confirm the malfunction and locate the faulty sensor.

[0013] Preferably, the degradation strategy library includes dedicated degradation strategies for combined abnormal scenarios; During the degradation control process, an anomaly propagation suppression mechanism is implemented to monitor the risk of anomalies spreading to the core control unit and to reduce control authority in advance. Simultaneously, an anomaly log is recorded, which includes at least the anomaly type, occurrence time, location result, anomaly level, degradation mode, and control parameters.

[0014] Preferably, when the anomaly classification result is a severe anomaly and / or when the vehicle attitude parameters are determined to exceed the preset stability boundary, an emergency stability degradation strategy is triggered. The emergency stabilization degradation strategy includes at least: outputting a forced intervention control command for auxiliary support, causing the auxiliary support actuators on the left and right sides of the vehicle to switch from a retracted state to an extended state, and adjusting the extension amount and / or support force of the auxiliary support actuators on both sides in an independent closed-loop manner; At the same time, the vehicle attitude target is locked to the preset safe attitude, and the drive control output and / or steering control output are limited to the preset safe range until the vehicle attitude parameters are restored to within the stable boundary or the vehicle enters the parking stability mode.

[0015] Preferably, when the anomaly source location result is an actuator anomaly, the control authority allocation scheme includes a control authority convergence and takeover compensation strategy for the abnormal actuator; The takeover compensation strategy includes at least one of the following or a combination thereof: a) When the steering actuator malfunctions, lock the current steering angle or return the steering angle to a preset safe angle, and compensate for vehicle attitude deviation based on the coordinated adjustment of drive control output and auxiliary support control output. b) When the drive actuator malfunctions, the drive control output is limited or cut off, the brake control output is linked to make the vehicle decelerate smoothly, and the auxiliary support actuator is triggered to enter the parking stability mode. c) When the auxiliary support actuator malfunctions, limit the drive control output and steering control output, and switch the vehicle control mode to low-speed safety mode; Furthermore, during the execution of the takeover compensation strategy, the scheduling of control commands follows the priority rule that stability control takes precedence over steering control, and steering control takes precedence over drive control.

[0016] Compared with the prior art, the beneficial effects of the present invention are: 1. To address the shortcomings of existing two-wheeled vehicle anomaly detection methods, such as single-dimensionality detection, lack of cross-dimensional logic verification, poor adaptability of degradation strategies to combined anomalies, lack of targeted takeover compensation schemes for actuator anomalies, lack of horizontal and vertical collaborative design in stability control, and reliance on mechanical gyro torque stabilization devices which increases hardware costs, this invention achieves full-dimensional detection and precise localization of anomalies in sensors, communication links, and actuators without relying on mechanical gyro torque stabilization devices. It combines anomaly level and type matching with a graded and scenario-based degradation strategy to construct a horizontally and vertically collaborative stability control model. The invention designs actuator anomaly takeover compensation strategies and emergency stability degradation strategies, and dynamically adjusts control parameters through adaptive data acquisition and feedback to achieve closed-loop optimization of the entire process from anomaly detection to degradation control to stability maintenance. This effectively addresses single and combined anomaly conditions, significantly improving vehicle controllability and driving safety under abnormal conditions. 2. This invention integrates multiple methods of preprocessing, including filtering and denoising, missing data completion, and dimension unification. Each method offers multiple options: filtering and denoising can choose from sliding window, low-pass, or other filtering methods depending on the actual interference type; missing value completion can use interpolation or historical estimation; and dimension unification can choose normalization or standardization. This approach demonstrates strong adaptability to various scenarios. Through this multi-method preprocessing, high-frequency noise in the collected data is effectively filtered out, missing data is completed to ensure data continuity, numerical differences between data with different dimensions are eliminated, and invalid data is simultaneously removed to obtain a standardized set of data to be detected. This significantly improves data quality, avoids subsequent anomaly detection biases caused by noise, missing values, and dimension inconsistencies, and significantly enhances the accuracy of subsequent anomaly detection. 3. This invention designs targeted detection methods for three types of detection objects: sensors, communication links, and actuators. Sensor data is detected using multiple methods such as threshold verification and trend analysis. Communication data is detected to identify link anomalies through frame loss and timeout detection. Actuator data is detected based on command-response residuals to determine operational anomalies, achieving accurate identification of different types of anomalies. Simultaneously, a cross-dimensional correlation verification mechanism is added to verify the logical consistency between the three types of data. This can identify potential anomalies that cannot be detected by single-dimensional detection. Redundant sensors can also be used to cross-verify suspected anomaly sensors, further confirming anomalies and locating failed sensors. This significantly reduces the false positive and false negative rates of anomaly detection, accurately outputs suspected anomaly results and specific anomaly types, and improves the targeting of the entire anomaly handling process. 4. This invention, through anomaly source localization using data tracing and / or fault tree analysis, can accurately trace the transmission path and associated units of abnormal data, pinpointing the specific location of the anomaly. This avoids the shortcomings of existing technologies that can only determine the anomaly type but cannot locate the specific anomaly source. Anomaly classification is not solely based on the degree of impact, but rather combines multiple dimensions such as the degree of impact on vehicle stability, controllability, and recoverability. It is clearly divided into three levels: minor, moderate, and severe. Severe anomalies are precisely defined as anomalies that may lead to vehicle instability, making the classification standard more scientific and more aligned with actual driving needs. This achieves precise anomaly source localization and scientific anomaly level classification, avoiding mismatches in degradation strategies caused by ambiguous localization or improper classification, and improving the rationality and targeting of degradation control. 5. This invention generates control commands by constructing a collaborative control model for lateral and longitudinal stability control. Laterally, it suppresses roll tendencies, and longitudinally, it achieves smooth deceleration, balancing vehicle attitude and speed stability. The model supports multiple control algorithms such as PID and MPC to adapt to different control requirements. In case of communication failure, it can switch to a redundant communication link, and the stability control message is set as the highest priority to preempt communication resources, ensuring reliable and timely transmission of commands. A dedicated takeover compensation strategy is designed for actuator failures, and an emergency stability degradation strategy is also formulated. Furthermore, it does not rely on mechanical gyro torque stabilization devices throughout the process, reducing hardware costs. Compared to existing technologies with single lateral or longitudinal control and no redundant communication guarantees, this invention achieves collaborative control for vehicle stability, ensuring the reliability of control command transmission. The actuator takeover compensation and emergency stabilization strategies enable the vehicle to effectively cope with severe abnormal conditions. The absence of mechanical devices makes the technical solution more hardware-adaptable and lower in cost. Attached Figure Description

[0017] Figure 1 The present invention provides an operational flow diagram of an anomaly detection and degraded stability control method for two-wheeled vehicles. Figure 1 ; Figure 2 The present invention provides an operational flow diagram of an anomaly detection and degraded stability control method for two-wheeled vehicles. Figure 2 ; Figure 3 The present invention provides an operational flow diagram of an anomaly detection and degraded stability control method for two-wheeled vehicles. Figure 3 ; Figure 4 The present invention provides an operational flow diagram of an anomaly detection and degraded stability control method for two-wheeled vehicles. Figure 4 . Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example Please see Figure 1-4 As shown, the present invention provides a technical solution comprising the following steps: S1. Synchronously collect sensor monitoring data, actuator operation data and communication data between various control units of the two-wheeled vehicle to complete the initial data collection; S2. Perform noise reduction, completion, and standardization preprocessing on the collected data to generate a qualified set of data to be tested; S3. Based on the preprocessed data, perform multi-dimensional anomaly identification on sensors, communication links, and actuators, and mark suspected anomalies and their preliminary types. S4. Based on the suspected abnormal results, accurately locate the source of the abnormality and classify it according to the degree of impact of the abnormality on driving safety. S5. Based on the anomaly classification results, match the preset degradation strategy library to determine the target degradation mode and control permission allocation scheme; S6. Based on the target degradation mode, construct a vehicle stability control model, output adapted degradation control commands, and drive actuator actions. S7. Real-time acquisition of vehicle status feedback data after downgrade control, and dynamic adjustment of control parameters; It is achieved without relying on mechanical gyroscope torque stabilization devices.

[0020] In this embodiment, the sensor monitoring data includes at least vehicle attitude parameters and motion state parameters; The actuator operating data includes at least the actuator output value and response delay; The vehicle network communication data includes message data and / or communication status data between the controller, sensors and actuators; And timestamp calibration is used to achieve synchronous aggregation of data from different sources.

[0021] Specifically, the vehicle attitude parameters are parameters characterizing the vehicle's spatial attitude, such as roll angle, pitch angle, and yaw rate during vehicle operation; the motion state parameters are parameters characterizing vehicle motion, such as vehicle speed, acceleration, steering angle, and wheel speed; the actuator output value is the actual action output parameter after the actuator receives the control command; the response delay is the time difference between the actuator receiving the control command and generating the actual action; the communication status data are parameters characterizing the status of the vehicle network communication link, such as communication signal strength, data transmission rate, packet loss rate, and CRC check result. Timestamp calibration is achieved by configuring a unified clock reference for each acquisition module, calibrating and correcting the local timestamps generated when each module acquires data, unifying the timestamps of all data to this clock reference, and simultaneously performing time interpolation alignment on asynchronously acquired multi-source data to achieve time synchronization of data from different acquisition frequencies and different acquisition modules. The synchronized and aggregated data is classified and stored according to data type, and metadata information such as acquisition module identifier, data type identifier, and calibrated timestamp is added to each type of data.

[0022] In this embodiment, the preprocessing includes denoising using at least one of sliding window filtering, low-pass filtering, Kalman filtering, or median filtering; Missing values ​​are filled using at least one of interpolation or historical estimation. Dimensional unification is achieved using at least one of normalization or standardization methods. The standardization formula is as follows: , In the formula, For standardized data, This is the original data. The mean of the original data sequence. denoted as the standard deviation of the original data sequence.

[0023] Specifically, filtering and denoising selects an appropriate filtering method based on the data type and noise characteristics. For sensor data with random high-frequency noise, sliding window filtering or median filtering can be used; for communication data with continuous interference noise, low-pass filtering can be used; and for high-precision attitude data with system noise and measurement noise, Kalman filtering can be used. Missing value completion selects an appropriate method based on the duration and continuity of the missing data. For short-term single-point data loss, interpolation completion is used, calculating the missing value through linear or nonlinear interpolation based on the valid data before and after the missing point. For long-term continuous data loss, historical estimation completion is used, estimating the missing data value based on the historical data patterns under the same operating conditions and the current data change trend. Dimensional unification is performed after data completion and denoising. Normalization maps the data to a preset fixed numerical range. Standardization converts the data into a distribution with mean and variance conforming to preset standards. During the processing, a separate processing model is established for each type of data, and all parameter information during the data processing is retained. The preprocessed set of data to be detected is divided into subsets of sensor data, actuator data, and vehicle network communication data. The data in each subset is arranged in order according to timestamps, while invalid data and outliers that exceed the preset reasonable value range are removed.

[0024] In this embodiment, the anomaly detection includes: Perform threshold verification, trend analysis, rate of change detection, and / or consistency detection on sensor data to identify sensor anomalies; Perform frame loss detection, timeout detection, CRC check failure detection, and / or communication delay detection on the vehicle network communication data to determine communication link anomalies; The actuator operation data is processed by executing instructions—response residual detection and / or response timing detection—to determine actuator anomalies, implemented as follows: , In the formula, For the amount of data change, For time-varying quantities, if the data exceeds the threshold range or the rate of change shows an unexpected abrupt change, it is marked as a suspected sensor anomaly; communication anomaly identification includes signal strength detection, transmission delay detection, and data integrity detection. It verifies whether the transmitted data is lost through a checksum. If the communication signal strength is lower than a preset threshold, the transmission delay is higher than a preset threshold, or there is a data integrity problem, it is marked as a suspected communication anomaly; actuator anomaly identification compares the deviation between the actual output parameters and the theoretical output parameters of the actuator. If the deviation continues to exceed the preset allowable range, it is marked as a suspected actuator anomaly; all suspected anomalies are marked with a preliminary anomaly type.

[0025] Specifically, threshold verification of sensor data compares real-time data with a preset normal value range; values ​​exceeding the range are marked as potentially abnormal. Trend analysis, by fitting the change curve of sensor data, determines whether the data change trend conforms to the physical laws of normal vehicle operation; deviations from the normal trend are marked as potentially abnormal. Rate of change detection calculates the amount of change in sensor data per unit time; changes exceeding a preset range are marked as potentially abnormal. Consistency detection compares data of the same physical quantity collected by multiple sensors; deviations between data exceeding a preset threshold are marked as potentially abnormal. Frame loss detection for vehicle network communication data counts the number of lost messages within a preset time period; values ​​exceeding a preset threshold are marked as potentially abnormal. Timeout detection determines whether the message transmission and response time exceeds a preset time limit; exceeding this limit is marked as potentially abnormal. Communication delay detection calculates the transmission time of a message from the sender to the receiver; transmission times exceeding a preset range are marked as potentially abnormal. The instruction-response residual detection of actuator operating data calculates the deviation between the actual output value of the actuator and the theoretical output value of the control instruction. If the deviation continues to exceed the preset range, it is marked as a suspected anomaly. The response timing detection judges whether the actual action timing of the actuator is consistent with the timing requirements of the control instruction. If the timing deviation exceeds the preset range, it is marked as a suspected anomaly. All suspected anomalies are marked with the corresponding anomaly type code according to the preset coding rules.

[0026] In this embodiment, the anomaly source localization is achieved based on data tracing and / or fault tree analysis; The anomaly classification is determined at least based on the degree of impact of the anomaly on vehicle stability, its controllability and / or recoverability, and severe anomalies correspond to anomalies that may lead to vehicle instability.

[0027] Specifically, data tracing traces data backwards along its transmission path and generation chain, starting from the collection point of suspected abnormal data. It sequentially checks the status of equipment and modules in each stage of data collection, transmission, and processing to pinpoint the specific stage and hardware unit that generated the abnormal data. Fault tree analysis uses the suspected abnormal result as the top event and decomposes the possible causes of the top event layer by layer according to the vehicle control logic and data flow relationship. It constructs a fault tree model of the abnormal cause and verifies the status of each bottom event in the fault tree to determine the specific cause and fault unit of the abnormality. Data tracing and fault tree analysis can be performed separately or in combination. When performed in combination, the data tracing results narrow down the scope of fault tree analysis, and the fault tree analysis results verify the conclusions of data tracing, thus achieving precise location of the abnormal source. When classifying anomalies, the impact of the anomaly on vehicle stability is quantitatively assessed to determine whether the anomaly affects the vehicle's core control functions and whether it will cause the vehicle's driving state to deviate from the preset range. The controllability assessment of the anomaly determines whether the impact of the anomaly can be suppressed and compensated through the vehicle's own control strategy. The recoverability assessment determines whether the anomaly is a temporary fault and whether it can be restored to normal through restarting, resetting, etc. The anomaly level is classified by combining the results of the three assessments. At the same time, clear judgment thresholds and judgment rules are set for each level of anomaly to form a standardized anomaly classification assessment system.

[0028] In this embodiment, the degradation strategy library includes multiple degradation modes, and different degradation modes correspond to different anomaly types and anomaly levels; Among them, minor anomalies correspond to monitoring mode; moderate anomalies correspond to speed limit and amplitude limit mode; and severe anomalies correspond to safe speed reduction or parking stability mode. Furthermore, in the event of severe anomalies, the control permission allocation scheme forcibly sets the control permissions related to attitude stabilization control to the highest priority and restricts the control permissions of non-core functions.

[0029] Specifically, the degradation strategy library is a set of strategies pre-built and stored in the vehicle controller. The library configures corresponding degradation modes for each combination of anomaly type and severity, and presets corresponding control parameter thresholds, actuator action rules, and control permission allocation rules for each degradation mode. In monitoring mode, the original vehicle control strategy remains unchanged; only the abnormal unit's status is continuously monitored and data is recorded without altering control parameters or permissions. In speed-limiting and amplitude-limiting mode, the maximum vehicle speed and maximum actuator output are limited based on the anomaly type and its impact range. The parameter configuration of core control functions is optimized, and some non-core auxiliary control functions are disabled. In safe deceleration or parking stability mode, the vehicle's active deceleration control logic is activated, controlling the vehicle to reduce its speed according to a preset deceleration curve. In parking stability mode, the vehicle is smoothly braked to a stop. The control permission allocation scheme prioritizes control functions based on their importance, with attitude stability control having the highest priority, followed by braking control and steering control, and finally drive control and various auxiliary controls. In cases of severe anomalies, the control command sending permissions for non-core functions are cut off, and the vehicle network's communication resources and actuator action resources are preferentially allocated to attitude stability control-related functional modules.

[0030] In this embodiment, the degraded stability control command is generated by constructing a collaborative control model of lateral stability control and longitudinal stability control; Lateral stability control includes adjusting steering control output and / or auxiliary support control output to suppress roll tendency; longitudinal stability control includes limiting drive control output and / or applying braking control output to achieve smooth deceleration. The collaborative control model employs at least one of proportional-integral-derivative (PID) control, model predictive control (MPC), or robust control. Furthermore, when an abnormality is detected in the communication link, the stability control command is switched to the redundant communication link for transmission. The network message of the stability control command has a higher priority than the drive and steering control messages, in order to preempt the vehicle network communication resources. The control model uses the PID control algorithm to calculate the core control commands, and the control formula is as follows: , In the formula, To control the output value of the command, This refers to the deviation between the actual state of the vehicle and the target stable state. This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients; After the control command is generated, it is first transmitted through a communication link that has been confirmed to be free of abnormalities. If a communication abnormality is found, it is immediately switched to a redundant communication link, and a CRC check code is added to ensure the integrity of the command transmission. After the actuator receives the command, it needs to provide feedback on the execution status within a preset response time. If the actuator is detected to have not acted according to the command, the control model will adjust the command parameters in real time.

[0031] Specifically, the collaborative control model for lateral and longitudinal stability control aims to maintain the vehicle's preset stable attitude and driving state. It uses vehicle attitude parameters and driving parameters as inputs, and steering control outputs, auxiliary support control outputs, drive control outputs, and braking control outputs as outputs, establishing a control logic relationship between inputs and outputs. In lateral stability control, the output torque and angle of the steering control are adjusted based on the vehicle's roll angle and yaw rate, while the action state of the auxiliary support actuator is adjusted according to the vehicle's roll trend. This synergistic action suppresses vehicle roll. In longitudinal stability control, the output power and torque of the drive motor are limited in stages based on the vehicle's current speed and deceleration requirements, while simultaneously adjusting the action state of the auxiliary support actuator according to the vehicle's roll trend. A preset braking curve applies braking pressure to achieve smooth, shock-free vehicle deceleration. The collaborative control model can employ multiple control algorithms individually or in combination. The model's control parameters are adaptively configured based on the vehicle's driving status and anomaly type. During model operation, real-time vehicle status data is collected as feedback to achieve closed-loop control. Communication link anomalies are detected in real-time by the vehicle network's status monitoring module. Upon detection of an anomaly, a communication link switching mechanism is immediately triggered, switching the stability control command transmission channel from the primary communication link to a redundant communication link. Simultaneously, the vehicle network's message scheduling protocol configures the highest priority identifier for stability control command network messages. When network communication resources are scarce, stability control command messages are transmitted first, ensuring timely and reliable transmission of commands.

[0032] In this embodiment, the status feedback data includes at least attitude parameters, vehicle speed parameters, actuator response parameters, and communication status parameters; The sampling frequency is adaptively adjusted according to the anomaly level; Based on the feedback data, deviation trend assessment and / or cumulative error assessment are performed. When the deviation exceeds the threshold, the degradation control parameters are updated. The collected feedback data is compared with the preset stability threshold range in the corresponding degradation mode to calculate the cumulative deviation between the actual state and the target stable state. ; If the cumulative deviation is within the allowable range, maintain the current downgraded control parameters; if the cumulative deviation exceeds the allowable range, make targeted adjustments based on the marked anomaly type: for sensing deviations caused by sensor anomalies, optimize the weights of the state estimation model to weaken the impact of abnormal sensor data.

[0033] Specifically, attitude parameters include real-time vehicle attitude data such as roll angle, pitch angle, and yaw rate during the degraded control process; vehicle speed parameters include real-time vehicle speed, acceleration, and deceleration; actuator response parameters include the actual output value, response delay, and action completion rate of each actuator after receiving the degraded stabilization control command; and communication status parameters include the signal strength, transmission rate, and packet loss rate of the main and redundant communication links of the vehicle network during the degraded control process. The status feedback data acquisition module shares hardware units with the previous data acquisition module. The acquisition frequency is preset with multiple adjustment standards according to the anomaly level. Minor anomalies use the basic acquisition frequency, moderate anomalies increase the acquisition frequency to several times the basic frequency, and severe anomalies use the highest acquisition frequency. The adjustment of the acquisition frequency is achieved by sending commands to each acquisition module through the vehicle controller, and each module responds in real time and adjusts its acquisition cycle. Deviation trend assessment compares the real-time collected feedback data with the preset steady-state target value, calculates the real-time deviation value, and fits the deviation change curve to determine whether the deviation trend is converging or diverging. Cumulative error assessment integrates the real-time deviation value over a preset time period to obtain the cumulative error value. Deviation trend assessment and cumulative error assessment can be performed individually or in combination. Corresponding thresholds are set for each assessment. When either assessment result exceeds the threshold, the vehicle controller immediately initiates the degraded control parameter update process. Based on the type, magnitude, and trend of the deviation, the control parameters of the cooperative control model, the action parameters of the actuators, and the output parameters of the control commands are adjusted accordingly. After the parameter update, it is sent to each execution module in real time to achieve dynamic optimization of degraded control.

[0034] In this embodiment, the anomaly detection further includes a cross-dimensional correlation verification mechanism, which identifies suspected anomalies by verifying the logical consistency between sensor data, communication data and actuator data; When a suspected sensor malfunction is detected, redundant sensors are invoked for cross-validation to confirm the malfunction and locate the faulty sensor.

[0035] Specifically, the cross-dimensional correlation verification mechanism is based on the physical laws and control logic of vehicle operation. It establishes a correlation verification model between sensor data, communication data, and actuator data, pre-setting normal logical relationships and deviation ranges between various types of data. For example, changes in steering angle sensor data should correspond logically to changes in yaw rate sensor data; drive control command message data should correspond logically to drive actuator output data; and communication link delay data should correspond logically to actuator response delay data. During the verification process, key feature values ​​of the three types of data are extracted in real time and substituted into the correlation verification model for verification. If the logical relationship between the data deviates from the preset range and the deviation continuously exceeds the threshold, then... Record the suspected anomaly, and record the correlation and deviation of the abnormal data. When a suspected sensor anomaly is detected, the vehicle controller sends a command to activate the corresponding redundant sensor, collect the same physical quantity data as the suspected anomaly sensor, and synchronously compare the data collected by the redundant sensor with the data collected by the suspected anomaly sensor to calculate the deviation value between the two. If the deviation value continues to exceed the preset cross-validation threshold, the sensor is confirmed to be abnormal. At the same time, combined with the sensor's hardware identifier and data acquisition link information, the failed sensor hardware unit is accurately located. If there are multiple redundant sensors, cross-validation is performed in sequence to improve the accuracy of anomaly judgment and avoid misjudgment caused by the failure of a single redundant sensor. In this embodiment, the degradation strategy library includes dedicated degradation strategies for combined abnormal scenarios; During the degradation control process, an anomaly propagation suppression mechanism is implemented to monitor the risk of anomalies spreading to the core control unit and to reduce control authority in advance. Simultaneously, an anomaly log is recorded, which includes at least the anomaly type, occurrence time, location result, anomaly level, degradation mode, and control parameters.

[0036] Specifically, combined anomaly scenarios refer to situations where two or more different types of anomalies occur simultaneously. Dedicated degradation strategies are pre-built for various common combined anomaly scenarios. Based on the level, type, and impact of each individual anomaly within the combined anomaly, the overall level of the combined anomaly is determined. Simultaneously, appropriate control strategies are formulated, clarifying the priority of each anomaly. Degradation modes and control parameters are configured according to the principles of suppressing core anomalies first, then handling secondary anomalies, and ensuring attitude stability first, then maintaining driving functions. The anomaly propagation suppression mechanism monitors the status data of each control unit and hardware module in real time through the vehicle controller, analyzes the associated modules and data transmission links of the anomaly unit, and determines whether the anomaly has a tendency to propagate to core control units such as the vehicle controller, attitude stability control module, and braking control module. Upon detecting a risk of spread, the control authority contraction process is immediately initiated, cutting off unnecessary data transmission links between the abnormal unit and the core control unit, restricting the abnormal unit's control command sending authority, prioritizing the allocation of core control unit resources to basic stability control functions, and simultaneously implementing state protection for the core control unit to prevent abnormal data and fault signals from interfering with it. The anomaly log adopts a real-time recording and cyclic storage method. In addition to basic recording content, it also includes information such as anomaly duration, anomaly handling measures, control parameter adjustment records, vehicle status change data, and actuator action records. The log data is arranged in order according to timestamps, and each log entry is given a unique identifier. It also supports local storage and external reading of logs, facilitating subsequent anomaly analysis and fault diagnosis.

[0037] In this embodiment, when the anomaly classification result is a severe anomaly and / or when the vehicle attitude parameters are determined to exceed the preset stability boundary, an emergency stability degradation strategy is triggered. The emergency stabilization degradation strategy includes at least: outputting a forced intervention control command for auxiliary support, causing the auxiliary support actuators on the left and right sides of the vehicle to switch from a retracted state to an extended state, and adjusting the extension amount and / or support force of the auxiliary support actuators on both sides in an independent closed-loop manner; At the same time, the vehicle attitude target is locked to the preset safe attitude, and the drive control output and / or steering control output are limited to the preset safe range until the vehicle attitude parameters are restored to within the stable boundary or the vehicle enters the parking stability mode.

[0038] Specifically, the preset stability boundary for vehicle attitude parameters is a threshold value pre-set based on the vehicle's structural characteristics and driving performance. When any attitude parameter exceeds this threshold, the vehicle is immediately determined to be in a state of attitude instability risk. This, along with severe abnormal results, serves as the trigger condition for the emergency stability degradation strategy. Upon triggering, the vehicle controller immediately switches the control mode to emergency stability control mode, prioritizing the execution of the emergency stability degradation strategy. A forced intervention control command for auxiliary support is sent to the controller of the auxiliary support actuator. Upon receiving the command, the controller drives the actuator's drive components to move, causing the auxiliary support mechanisms on both sides to move from the retracted position to the extended position. During the extension process, the position and force data of the two mechanisms are collected in real time. Using an independent closed-loop control method, the extension amount of the left and right auxiliary support actuators is adjusted according to the vehicle's real-time roll angle and yaw rate to maintain the vehicle body's stability. The system adjusts the support force of both sides of the mechanism according to the vehicle's weight and driving status to ensure the stability and reliability of the support and avoid secondary instability caused by excessive or insufficient support force on one side. After locking the vehicle's attitude target as the preset safe attitude, the parameters of this safe attitude are used as the sole control target of the collaborative control model. The action parameters of each actuator are continuously adjusted to restore the vehicle to this safe attitude. The preset safe range of the drive control output and steering control output is a parameter range that can ensure the basic stability of the vehicle and will not aggravate attitude instability. Within this range, small adjustments are made according to the vehicle's attitude recovery. Drive and steering actions beyond the range are strictly prohibited. When the vehicle's attitude parameters recover to within the stable boundary, the control range can be gradually relaxed. When the vehicle enters the parking stability mode, the auxiliary support actuator is kept in the deployed state until the emergency stability control mode is manually released.

[0039] In this embodiment, when the anomaly source location result is that the actuator is abnormal, the control authority allocation scheme includes a control authority convergence and takeover compensation strategy for the abnormal actuator. The takeover compensation strategy includes at least one of the following or a combination thereof: a) When the steering actuator malfunctions, lock the current steering angle or return the steering angle to a preset safe angle, and compensate for vehicle attitude deviation based on the coordinated adjustment of drive control output and auxiliary support control output. b) When the drive actuator malfunctions, the drive control output is limited or cut off, the brake control output is linked to make the vehicle decelerate smoothly, and the auxiliary support actuator is triggered to enter the parking stability mode. c) When the auxiliary support actuator malfunctions, limit the drive control output and steering control output, and switch the vehicle control mode to low-speed safety mode; Furthermore, during the execution of the takeover compensation strategy, the scheduling of control commands follows the priority rule that stability control takes precedence over steering control, and steering control takes precedence over drive control.

[0040] Working principle: The system collects vehicle attitude and motion parameters through various sensors mounted on the vehicle, collects operational data such as actuator output values ​​and response delays through the actuator controller, and collects message data and communication status data between the controller, sensors, and actuators through the onboard network module, covering three core dimensions: vehicle driving status, actuator working status, and control unit communication status. At the same time, a timestamp calibration mechanism is used to map various types of data from different collection frequencies and data sources to the same time axis, realizing the time synchronization and collection of multi-source data, ensuring the consistency of data sequence, and providing a complete and synchronous raw data foundation for subsequent data processing and anomaly detection. Standardization improves data quality, eliminates various data interferences, and adapts to the data analysis needs of subsequent anomaly detection. Based on the actual situation of the original data, at least one of the following preprocessing methods is selected: filtering and denoising, missing data completion, and standardization. High-frequency interference signals can be filtered using methods such as sliding window filtering and low-pass filtering; missing data items can be filled using interpolation or historical estimation; and normalization or standardization can unify the dimensions of data with different units and value ranges. Through these preprocessing operations, invalid interference information is eliminated, ensuring the continuity, validity, and consistency of the data, ultimately forming a standardized set of data to be detected. This preprocessing is applied to three types of sensors, communication links, and actuators. The system employs differentiated identification methods to accurately screen all types of anomalies. Sensor data is assessed using threshold verification and trend analysis to determine if it exceeds normal ranges or exhibits unexpected changes. Vehicle network communication data is analyzed using frame loss detection and timeout detection to identify link transmission faults. Actuator operation data is analyzed using command-response residual detection to determine if the executed actions match the commands. A cross-dimensional correlation verification mechanism is introduced to verify the logical consistency between sensor, communication, and actuator data, identifying potential anomalies that cannot be detected by single-dimensional detection. If a suspected sensor anomaly is detected, redundant sensors are used for cross-verification to further confirm the abnormal state. Finally, the detected problems are categorized and labeled. Output suspected anomaly results and corresponding anomaly types; employ data tracing and / or fault tree analysis to trace the transmission path, associated control units, and logical relationships of suspected anomaly data, accurately locating the specific hardware unit or software module where the anomaly occurred, and clarifying the root cause of the anomaly; simultaneously, based on the degree of impact of the anomaly on vehicle stability, the vehicle's controllability of the anomaly, and the anomaly's recoverability, classify the anomaly into three levels: minor anomaly, moderate anomaly, and severe anomaly, with severe anomalies explicitly defined as anomalies that may lead to vehicle instability; this step completes the precise and quantitative judgment of anomalies, forming clear anomaly source location results and anomaly classification conclusions; based on the anomaly classification and location results, achieve precise matching and control of degradation strategies. Reasonable allocation of permissions clarifies the core direction of subsequent stability control; the preset degradation strategy library contains exclusive degradation strategies adapted to different anomaly types, anomaly levels, and combined anomaly scenarios. According to the rules of matching monitoring mode for minor anomalies, speed-limiting and amplitude-limiting mode for moderate anomalies, and safe speed reduction or parking stability mode for severe anomalies, the target degradation mode corresponding to the current anomaly is determined; at the same time, a corresponding control permission allocation scheme is generated, setting attitude stability control-related permissions to the highest priority under severe anomalies, and restricting control permissions for non-core functions; during the process of degradation strategy matching and permission allocation, an anomaly propagation suppression mechanism will also be activated to monitor the risk of anomalies spreading to the core control unit and shrink control permissions in advance to prevent the risk of secondary anomalies;Reliable degraded stability control commands are generated by constructing a collaborative control model to drive the actuators to complete corresponding control actions. First, a collaborative control model for lateral and longitudinal stability control is constructed. PID control, model predictive control, and other algorithms can be selected to optimize the control logic. Lateral stability control suppresses vehicle roll by adjusting steering control output and / or auxiliary support control output, while longitudinal stability control achieves smooth vehicle deceleration by limiting drive control output and / or applying braking control output. After the control commands are generated, if a communication link anomaly is detected, the stability control commands are immediately switched to a redundant communication link for transmission, and the network message priority of the stability control commands is set higher than that of drive and steering control messages to preempt onboard network communication resources and ensure reliable command transmission. If a severe anomaly or vehicle attitude exceeds the stability boundary, an emergency stability degradation strategy is triggered, controlling the auxiliary support actuators to extend... The system independently adjusts the support status; if the actuator malfunctions, a corresponding takeover compensation strategy will be implemented, and control command scheduling follows the rule of prioritizing stability control over steering control, and steering control over drive control; through real-time status feedback and dynamic parameter adjustment, the system continuously ensures the vehicle's stable driving state in degraded mode; firstly, multi-dimensional status feedback data after vehicle degraded control is collected, covering at least attitude parameters, vehicle speed parameters, actuator response parameters, and communication status parameters, and the collection frequency is adaptively adjusted according to the anomaly level, increasing the collection frequency for moderate and severe anomalies to ensure timely capture of subtle changes in vehicle status; subsequently, based on the collected feedback data, deviation trend assessment and / or cumulative error assessment are conducted, comparing the actual status data with the vehicle's preset stable state data. If the assessment finds that the deviation exceeds a preset threshold, the degraded control parameters are immediately dynamically adjusted to optimize the control command output.

[0041] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their likenesses.

[0042] The present invention and its embodiments have been described above. This description is not restrictive, and the accompanying drawings are only one embodiment of the present invention; the actual structure is not limited thereto. In conclusion, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the invention, such designs should fall within the protection scope of the present invention.

Claims

1. A method for anomaly detection and degraded stability control of two-wheeled vehicles, characterized in that, Includes the following steps: S1. Synchronously collect sensor monitoring data, actuator operation data, and vehicle network communication data of the two-wheeled vehicles; S2. Preprocess the collected data to obtain a set of data to be detected. The preprocessing includes at least one of filtering and noise reduction, missing data completion, and standardization. S3. Based on the dataset to be detected, perform anomaly detection on sensors, communication links and actuators, and output suspected anomaly results and anomaly types; S4. Based on the suspected abnormal results, locate the abnormal source and classify the abnormality. The abnormality classification includes at least minor abnormality, moderate abnormality and severe abnormality. S5. Based on the anomaly classification results, match the preset degradation strategy library, determine the target degradation mode, and generate a control permission allocation scheme; S6. Output a degradation stabilization control command according to the target degradation mode. The degradation stabilization control command includes at least one of the following: drive control command, steering control command, braking control command and auxiliary support control command. S7. Collect vehicle status feedback data after degradation control, and dynamically adjust degradation control parameters based on the feedback data to maintain the vehicle in a preset stable state.

2. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The sensor monitoring data includes at least vehicle attitude parameters and motion state parameters; The actuator operating data includes at least the actuator output value and response delay; The vehicle network communication data includes message data and / or communication status data between the controller, sensors and actuators; And timestamp calibration is used to achieve synchronous aggregation of data from different sources.

3. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The preprocessing includes denoising using at least one of sliding window filtering, low-pass filtering, Kalman filtering, or median filtering. Missing values ​​are filled using at least one of interpolation or historical estimation. Use at least one of the normalization or standardization methods to unify the dimensions.

4. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The anomaly detection includes: Perform threshold verification, trend analysis, rate of change detection, and / or consistency detection on sensor data to identify sensor anomalies; Perform frame loss detection, timeout detection, CRC check failure detection, and / or communication delay detection on the vehicle network communication data to determine communication link anomalies; Execute instructions on the actuator operating data—response residual detection and / or response timing detection—to determine actuator anomalies.

5. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The anomaly source localization is achieved based on data tracing and / or fault tree analysis; The anomaly classification is determined at least based on the degree of impact of the anomaly on vehicle stability, its controllability and / or recoverability, and severe anomalies correspond to anomalies that may lead to vehicle instability.

6. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The degradation strategy library includes multiple degradation modes, with different degradation modes corresponding to different exception types and exception levels; Among them, minor anomalies correspond to monitoring mode; moderate anomalies correspond to speed limit and amplitude limit mode; and severe anomalies correspond to safe speed reduction or parking stability mode. Furthermore, in the event of severe anomalies, the control permission allocation scheme forcibly sets the control permissions related to attitude stabilization control to the highest priority and restricts the control permissions of non-core functions.

7. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The downgraded stability control command is generated by constructing a collaborative control model of horizontal stability control and vertical stability control; Lateral stability control includes adjusting steering control output and / or auxiliary support control output to suppress roll tendency; longitudinal stability control includes limiting drive control output and / or applying braking control output to achieve smooth deceleration. The collaborative control model employs at least one of proportional-integral-derivative (PID) control, model predictive control (MPC), or robust control. Furthermore, when an anomaly in the communication link is detected, the stability control command is switched to the redundant communication link for transmission. The network message of the stability control command has a higher priority than the drive and steering control messages, and is used to preempt the vehicle network communication resources.

8. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The status feedback data includes at least attitude parameters, vehicle speed parameters, actuator response parameters, and communication status parameters; The sampling frequency is adaptively adjusted according to the anomaly level; Based on the feedback data, deviation trend assessment and / or cumulative error assessment are performed, and the degradation control parameters are updated when the deviation exceeds the threshold.

9. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The anomaly detection further includes a cross-dimensional correlation verification mechanism, which identifies suspected anomalies by verifying the logical consistency between sensor data, communication data and actuator data. When a suspected sensor malfunction is detected, redundant sensors are invoked for cross-validation to confirm the malfunction and locate the faulty sensor.

10. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: The degradation strategy library includes dedicated degradation strategies for combined abnormal scenarios; During the degradation control process, an anomaly propagation suppression mechanism is implemented to monitor the risk of anomalies spreading to the core control unit and to reduce control authority in advance. Simultaneously, an anomaly log is recorded, which includes at least the anomaly type, occurrence time, location result, anomaly level, degradation mode, and control parameters.

11. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: When the anomaly classification result is a severe anomaly and / or when the vehicle attitude parameters are determined to exceed the preset stability boundary, an emergency stability degradation strategy is triggered. The emergency stabilization degradation strategy includes at least: outputting a forced intervention control command for auxiliary support, causing the auxiliary support actuators on the left and right sides of the vehicle to switch from a retracted state to an extended state, and adjusting the extension amount and / or support force of the auxiliary support actuators on both sides in an independent closed-loop manner; At the same time, the vehicle attitude target is locked to the preset safe attitude, and the drive control output and / or steering control output are limited to the preset safe range until the vehicle attitude parameters are restored to within the stable boundary or the vehicle enters the parking stability mode.

12. The anomaly detection and degraded stability control method for a two-wheeled vehicle according to claim 1, characterized in that: When the anomaly source location result is that the actuator is abnormal, the control authority allocation scheme includes a control authority convergence and takeover compensation strategy for the abnormal actuator. The takeover compensation strategy includes at least one of the following or a combination thereof: a) When the steering actuator malfunctions, lock the current steering angle or return the steering angle to a preset safe angle, and compensate for vehicle attitude deviation based on the coordinated adjustment of drive control output and auxiliary support control output. b) When the drive actuator malfunctions, the drive control output is limited or cut off, the brake control output is linked to make the vehicle decelerate smoothly, and the auxiliary support actuator is triggered to enter the parking stability mode. c) When the auxiliary support actuator malfunctions, limit the drive control output and steering control output, and switch the vehicle control mode to low-speed safety mode; Furthermore, during the execution of the takeover compensation strategy, the scheduling of control commands follows the priority rule that stability control takes precedence over steering control, and steering control takes precedence over drive control.

Images