A differential steering safe parking control method for a vehicle in a severe steering failure scenario

Abstract

This invention discloses a differential steering safety stopping control method for vehicles in severe steering failure scenarios. The method establishes a residual steering capability envelope model considering the coupling effect of longitudinal vehicle speed and acceleration / deceleration. Secondly, it constructs a vehicle motion error boundary to cover trajectory deviations caused by model mismatch, external disturbances, and faults. Thirdly, it constructs a safe stopping trajectory and collision avoidance constraints for dynamic or static obstacles. Finally, in the online safe stopping control module, the residual steering capability constraints and robust safety boundaries are introduced into the safe stopping optimization control framework to solve for the optimal safe stopping trajectory and execute control commands. Combined with error feedback control, this ensures that the vehicle trajectory remains within the target trajectory safety boundary. This invention can ensure that, under severe steering failure conditions, the vehicle fully utilizes its remaining differential steering capability to safely move the vehicle out of the active lane and complete the stop, while ensuring kinematic feasibility and robust collision avoidance safety.

Description

Technical Field

[0001] This invention relates to the field of intelligent vehicle control technology, and in particular to a differential steering safety parking control method for a vehicle steering failure scenario. Background Technology

[0002] With the rapid development of advanced autonomous driving technology (Level 4 and above), the vehicle's safety assurance capability after system failure (Fail-Operational) has become a core research direction. Among these, the steering system, as a key actuator controlling the vehicle's lateral movement, directly threatens driving safety in the event of a serious malfunction. Advanced autonomous driving requires the ability to automatically achieve minimal risk, including moving the vehicle out of active lanes.

[0003] In existing technical solutions, responses to steering failure primarily rely on hardware redundancy or simple lane-keeping emergency braking. However, in high-speed driving or complex traffic flow, simply stopping in the lane can easily lead to secondary accidents such as rear-end collisions. Safely moving the vehicle out of the active lane is a more ideal emergency strategy. When the main steering system fails severely (e.g., the steering motor seizes up or completely loses power), the vehicle must rely on the yaw moment generated by the four-wheel independent drive or braking system (i.e., differential steering) to maintain lateral control. However, the mechanical mechanism of differential steering is fundamentally different from traditional active front-wheel steering. The yaw moment generated by differential steering is strongly coupled with the vehicle's longitudinal acceleration and deceleration (limited by the tire friction circle). Traditional path planning and control methods based on active steering models do not fully consider this coupling limitation, and the generated trajectory often exceeds the vehicle's actual physical limits under failure conditions, leading to vehicle instability.

[0004] Meanwhile, under fault conditions, due to the complex dynamic characteristics of the differential actuator, the error of the vehicle's dynamic model increases significantly, and it is easily affected by random disturbances such as road surface unevenness. Traditional collision avoidance algorithms based on nominal models cannot guarantee safety during actual driving through simple geometric constraints. Summary of the Invention

[0005] This invention aims to solve the problem that traditional path tracking and obstacle avoidance fail when a vehicle's steering ability is significantly reduced due to a serious failure in the steering system, caused by differences in differential steering dynamics and model uncertainties.

[0006] To achieve the above objectives, a first aspect of the present invention provides a differential steering safety parking control system for a severe vehicle steering failure scenario, comprising: a sensor array, a fault identification unit, an offline fault safety boundary calculation module, an online safety parking control module, and an actuator subsystem, wherein:

[0007] The sensor array is responsible for collecting vehicle dynamics data in real time, fusing environmental data using visual and radar algorithms, and finally sending the processed vehicle motion state and environmental perception information to the fault identification unit and online safe parking control module in real time. The fault identification unit is used to accurately identify the specific failure location of the steering system, drive system or braking system by analyzing the response residual, current feedback or hydraulic state and vehicle dynamic state of the actuator fed back by the sensor combination, and further calculate the remaining efficiency ratio and asymmetric fault characteristic parameters of the failed actuator, thereby generating a vehicle actuator fault state feedback signal containing the fault location and positive and negative remaining efficiency coefficients, and sending it to the offline fault safety boundary calculation module and the online safe parking control module in real time. The offline fault safety boundary calculation module is used to receive fault feedback information from the fault identification unit and provide the online safe parking control module with error safety boundary, error feedback control matrix and residual steering capability signal; The online safe parking control module is used to receive sensing information from the sensor combination, execution fault status information from the fault identification unit, and error safety boundary, error feedback matrix, and residual steering capability from the offline fault safety boundary calculation module. Under severe steering failure conditions, it plans and tracks a safe parking trajectory to achieve safe parking and collision avoidance. The actuator subsystem includes a four-wheel independent braking brake system, a four-wheel independent drive system, a steer-by-wire system, and sensor combinations of the corresponding subsystems, used to execute the generated actuator subsystem execution commands.

[0008] Optionally, the offline fault-safe boundary calculation module includes: The error safety boundary calculation unit, based on the pre-calibrated model error distribution, calculates the convergence range of the vehicle's state error under closed-loop control for the current fault mode, and generates the error safety boundary and the corresponding error feedback matrix, which are used to quantify the maximum trajectory deviation of the vehicle under faults and disturbances. The vehicle residual steering capability calculation unit receives fault information, vehicle speed information, and acceleration status from the fault identification unit. Based on the vehicle's nonlinear dynamics model, it solves the optimization problem of the maximum yaw rate under the remaining actuator execution capability, obtains the residual maximum allowable yaw rate, and quickly retrieves the current available maximum yaw rate range of the vehicle using a parametric surface fitting model or a lookup table method.

[0009] Optionally, the online safe parking control module includes a safe parking target planning unit, a safe collision avoidance calculation unit, a parking trajectory optimization unit, a robust feedback control unit, and an execution command allocation unit, wherein: The safe parking planning unit searches for and determines the safe parking target point and the drivable area boundary, taking into account the error safety boundary, based on the environmental perception information and the current state of the vehicle, and sends them to the parking trajectory optimization unit. The safety collision avoidance calculation unit is used to perform geometric modeling of surrounding stationary and moving obstacles, use a convex polygon two-dimensional envelope to surround the faulty vehicle and surrounding obstacles, and use a smoothing function to process the directed distance constraint between the obstacle and the vehicle contour to generate a differentiable obstacle collision avoidance constraint that considers the safety boundary of motion error, and send it to the parking trajectory optimization unit. The parking trajectory optimization unit is used to integrate the vehicle's residual steering ability, error safety boundary, collision avoidance constraints, parking target, feasible area and vehicle dynamics model into a nonlinear optimization problem based on the optimal control framework. Under the premise of satisfying all constraints, it plans the optimal target trajectory from the current position to the safe parking point and the corresponding nominal control command, ensuring that the vehicle parking trajectory is safe and dynamically feasible, and sends the trajectory and nominal control command to the robust feedback control unit. The robust feedback control unit receives the target trajectory and nominal control commands and calculates the deviation between the actual vehicle state and the target trajectory. Using the error feedback matrix obtained offline, it calculates robust feedback control commands in real time to offset disturbances and model errors, ensuring that the actual vehicle trajectory always remains within the error safety boundary. The execution command allocation unit is used to receive actuator control commands from the robust feedback control unit, generate the final front wheel steering angle command, four-wheel braking torque command, and four-wheel drive torque command, and send them to the actuator subsystem.

[0010] To achieve the above objectives, a second aspect of the present invention provides a differential steering safety stopping control method for a vehicle steering failure scenario, comprising: Acquire vehicle dynamics status, fault information and environmental information, establish a vehicle path tracking dynamics model, and determine the remaining execution capacity and allowable execution range of each actuator based on actuator fault information; Based on the fault information, the vehicle state error safety boundary, the control input error safety boundary, and the corresponding state feedback matrix are calculated using the robust control invariant set. Based on the fault information, a residual steering capability envelope model is constructed, and the steering capability constraint range of the vehicle under fault conditions is determined. Determine safe parking areas and feasible areas based on vehicle status based on environmental information; Predict the trajectory of obstacles, calculate the directed distance between the vehicle and the obstacle, and construct collision avoidance constraints; Based on the vehicle dynamics model, residual steering capability constraints, error safety boundary and collision avoidance constraints, a nonlinear optimization problem is constructed and solved to obtain the nominal state trajectory and nominal control input. Based on the deviation between the actual vehicle state and the nominal trajectory, robust execution control input is calculated using the state feedback matrix. The vehicle actuators are controlled to execute corresponding control commands based on the control input.

[0011] Optionally, acquire vehicle dynamics state, fault information, and environmental information; establish a vehicle path tracking dynamics model; and determine the remaining execution capacity and permissible execution range of each actuator based on actuator fault information, including: Modeling is performed for vehicle path tracking dynamics that consider differential motion:

[0012]

[0013] In the formula, This represents the vehicle's longitudinal position in the geodetic coordinate system. This represents the horizontal position in the geodetic coordinate system. For the longitudinal speed of the vehicle, The longitudinal acceleration of the vehicle; The yaw angle of the vehicle. Let yaw rate be the vehicle's angular velocity. The yaw acceleration of the vehicle; The lateral speed of the vehicle. This refers to the vehicle's lateral acceleration. Let Z be the vehicle's moment of inertia along the z-axis. For the overall vehicle weight; This refers to the front wheel steering angle. ( (representing the front and rear axles respectively) represents the lateral forces on the front and rear axles. For the longitudinal forces on the front and rear axles, For four-wheel drive / braking force command ( (These represent the front left, front right, rear left, and rear right wheels, respectively). For four-wheel drive braking torque, and , The radius of the wheel; and These are the distances from the front axle to the center of gravity and the rear axle to the center of gravity, respectively. The wheelbase of the vehicle; This refers to the actual yaw moment generated by the four-wheel differential. For modeling errors or unknown disturbances; system state Control input ; The above system can be expressed as a nonlinear affine system:

[0014] in, Let be the system's autonomous function matrix. The system control input gain function matrix; The current status of the vehicle is obtained through a combination of sensors. And calculate the information of fixed obstacles or dynamic obstacles as This represents the geometric information of the obstacle and its positional relationship over time. A fixed obstacle is assumed to be... Its position does not change over time. , Represents the number of obstacles; Based on the actuator status feedback from the fault identification unit, the remaining actuator capability under vehicle execution failure conditions is calculated, and the following model is performed based on the fault information:

[0015]

[0016] in, , This is a matrix of positive and negative fault information for vehicle actuators. These are the positive residual actuator capacity coefficients for the steering system and the four-wheel drive / braking system, respectively, representing the ratio of residual actuator capacity to actuator capacity under fault-free conditions; The negative residual actuator capacity coefficient of the execution system; Let be a function that generates a matrix based on its diagonal elements. Then the permissible execution range of the vehicle actuator is:

[0017] in, The execution command for the i-th executor ( These represent the front wheel steering angle command and the left front, right front, left rear, and right rear wheel braking commands, respectively. The maximum and minimum values ​​for the i-th execution subsystem under healthy conditions.

[0018] Optionally, based on the fault information, the vehicle state error safety boundary, the control input error safety boundary, and the corresponding state feedback matrix are calculated using robust control invariant sets, including: Based on the modeling and estimation of model uncertainties, fault disturbances, and external disturbances under fault conditions, the error safety boundary of the vehicle state is obtained through robust control invariant set calculation. Error safety boundary with vehicle control input :

[0019] in, System status With open-loop trajectory status of autonomous driving error; For system control input With open-loop trajectory status input for autonomous driving error; This is the upper bound of the norm of the error boundary; The infinite norm of the variable; Simultaneously, the state feedback matrix under specific fault conditions is obtained. This matrix represents the vehicle state. function.

[0020] Optionally, a residual steering capability envelope model is constructed based on the fault information, and the steering capability constraint range of the vehicle under fault conditions is determined, including: The following optimization problem is constructed by combining differential vehicle path tracking dynamics:

[0021] In the formula, , These are the minimum and maximum values ​​of the front wheel steering angle that can be executed under fault-free conditions; , These represent the maximum braking and driving values ​​that can be performed on the four wheels under fault-free conditions, respectively. The road adhesion coefficient; The vertical load is for the four wheels; The acceleration condition is set. Under specific fault conditions, the speed With acceleration The above optimization problem is solved by iterating through the data to obtain the operating point with the maximum yaw rate. A cubic polynomial surface fitting is then performed on the resulting surface. The maximum yaw rate of the vehicle is obtained by using two polynomial surfaces and taking the minimum value.

[0022] in, This represents a surface fitted by a cubic polynomial. These are the fitting coefficients; Based on the symmetry and steering symmetry of the vehicle model, the minimum values ​​of vehicle negative acceleration and vehicle yaw rate are obtained. ; Calculate the quantitative envelope constraint range of residual steering capability under vehicle fault conditions. for:

[0023] The solution is complete.

[0024] Optionally, safe parking areas and feasible areas based on vehicle status can be determined, including: Based on high-precision maps or real-time sensing results, a convex polygon region that meets the vehicle's geometric dimensions is searched within the right shoulder of the road or emergency lane, and defined as the terminal safety set. ; Define the feasible area for vehicle status based on road boundaries and vehicle status restrictions. .

[0025] Optionally, predict the obstacle's trajectory, calculate the directed distance between the vehicle and the obstacle, and construct collision avoidance constraints, including: Calculate the faulty vehicle using the directed distance of a convex polygon. The directed distance to the obstacle, where the convex polygon The o-th vertex To convex polygon Two clockwise vertices and The defined boundary of the e-th hyperplane distance The calculation is as follows:

[0026]

[0027] in, Given a rotation matrix, Let be the quadratic norm of the variable; The approximate formula is:

[0028] in, It is an adjustable parameter. and The operations of approximating the maximum and minimum values ​​can be performed respectively. For function parameters, For parameter dimensions; The envelope of the faulty vehicle and obstacles Directed distance It can be approximated as:

[0029]

[0030]

[0031]

[0032] in, , , , ; and Convex polygons The number of vertices; The collision avoidance constraints for the vehicle are designed as follows:

[0033] in, This is a safety margin for vehicle collisions. Minkowski sum for convex sets.

[0034] Optionally, based on the vehicle dynamics model, residual steering capability constraints, error safety boundaries, and collision avoidance constraints, a nonlinear optimization problem is constructed and solved to obtain the nominal state trajectory and nominal control input, including: This problem integrates the parking objective, residual vehicle steering capability, error safety boundary, collision avoidance constraints, and vehicle dynamics model into a nonlinear parking optimization control problem in the variable time domain. The optimization objective is to generate a nominal trajectory sequence that enables a fast and safe stop. and nominal control input sequence The optimization problem is:

[0035] in, For the system discrete time, To optimize the discrete time of the optimal system, Optimize variables for system state. For system control optimization variables, The yaw rate is the optimization variable in the system state. , The dimension is the discrete state. Weighted by parking time; The permissible range of control commands to be executed after a vehicle malfunction; For the initial optimized state of the vehicle, corresponding exist Time value, When the vehicle is parked, the corresponding exist Time value; The current vehicle status; Minkowski difference for convex sets; To optimize performance metrics, it can be defined as follows: ; The above problem is solved using a nonlinear programming solver to obtain the optimal nominal control input sequence. and vehicle status trajectory .

[0036] To achieve the above objectives, a third aspect of this application provides an electronic device, including a processor and a memory; wherein the processor runs a program corresponding to the executable program code stored in the memory to implement the method described in the first aspect.

[0037] To achieve the above objectives, a fourth aspect of this application provides a non-transitory computer-readable storage medium having a computer program stored thereon that, when executed by a processor, implements the method described in the first aspect.

[0038] The embodiments of this invention have the following beneficial effects: For extreme conditions where autonomous vehicles experience severe steering failure, this invention effectively solves the collision avoidance safety problem caused by sudden changes in vehicle dynamics after a failure and the problem of insufficient real-time computation by constructing an architecture that combines offline fault-safe boundary calculation with online safe parking control. This invention utilizes parametric surface fitting to realize the vehicle's residual steering capability, and on this basis, constructs a variable time-domain nonlinear optimal control problem that includes safety error boundaries, smooth collision avoidance constraints, and residual steering capability constraints. Through the synergistic effect of nominal trajectory planning and robust feedback control, this invention can ensure that when the vehicle loses its main steering capability, it can fully utilize differential drive / braking potential, achieving robust and accurate autonomous obstacle avoidance and safe parking under the premise of strictly satisfying dynamic constraints and physical safety boundaries. Attached Figure Description

[0039] The above-described and additional aspects and advantages of the present invention will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, in which: Figure 1 A schematic diagram of the framework of a differential steering safety parking control system for a severe vehicle steering failure scenario provided in an embodiment of the present invention; Figure 2 A flowchart of a differential steering safety stop control method for a vehicle steering failure scenario provided by an embodiment of the present invention; Figure 3 This is a schematic diagram of the envelope fitting of the residual steering capability of a vehicle after severe steering failure, provided as an embodiment of the present invention. Detailed Implementation

[0040] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other. The present invention will now be described in detail with reference to the accompanying drawings and embodiments.

[0041] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. 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 should fall within the scope of protection of the present invention.

[0042] The following description, with reference to the accompanying drawings, describes a differential steering safety parking control system and method for a vehicle steering failure scenario according to an embodiment of the present invention.

[0043] Example 1 This embodiment provides a differential steering safety parking control system for scenarios of severe vehicle steering failure. For example... Figure 1 As shown, the system includes: a sensor array, a fault identification unit, an offline fault safety boundary calculation module, an online safe parking control module, and an actuator subsystem.

[0044] Specifically: (1) Sensor combination, used to perceive the vehicle's motion state and surrounding environment information in all directions, specifically includes high-precision hardware devices such as lidar, camera, integrated inertial navigation system (IMU) and wheel speed sensor. This module is responsible for collecting dynamic state data such as vehicle yaw rate, lateral acceleration, heading angle and wheel speed in real time, and using vision and radar algorithms to fuse perception of environmental data such as lane line curvature, road boundary geometric features and the position of surrounding vehicles or obstacles. Finally, the processed vehicle motion state and environmental perception information are sent to the fault identification unit and online safe parking control module in real time to provide data support for subsequent fault diagnosis, parking decision and trajectory planning.

[0045] (2) Fault identification unit, which monitors the health status of the vehicle's underlying actuators in real time and quantifies the degree of failure. This module accurately identifies the specific failure location of the steering system, drive system or braking system by analyzing the response residuals, current feedback or hydraulic status and vehicle dynamics status of the actuators fed back by the sensor combination, and further calculates the remaining efficiency ratio and asymmetric fault characteristic parameters of the failed actuator, thereby generating a vehicle actuator fault status feedback signal containing the fault location and positive and negative remaining efficiency coefficients, and sending it to the offline fault safety boundary calculation module and the online safe parking control module in real time.

[0046] (3) Offline Fault Safety Boundary Calculation Module: This module is configured to quickly provide vehicle performance limits and safety constraint parameters based on pre-calculated or offline fitting after a fault occurs. This module receives fault feedback information from the fault identification unit and provides the online safe parking control module with error safety boundaries, error feedback control matrices, and residual steering capability signals. Specifically, it includes an error safety boundary calculation unit and a vehicle residual steering capability calculation unit. Due to the diverse fault modes and the complexity of calculating safety boundaries and residual steering capabilities, this module obtains and stores the error safety boundaries and residual steering capabilities under different fault conditions through offline calculation. After obtaining fault information, it queries or interpolates relevant output signals based on the fault state feedback signals to reduce real-time and computational pressure.

[0047] In one embodiment of the present invention, the error safety boundary calculation unit calculates the convergence range of the vehicle's state error under closed-loop control based on a pre-calibrated model error distribution for the current fault mode, and generates an error safety boundary (and a corresponding error feedback matrix) to quantify the maximum trajectory deviation of the vehicle under faults and disturbances. The calculation results for different fault conditions are performed offline and stored to reduce real-time and computational pressure.

[0048] In addition, the vehicle residual steering capability calculation unit receives fault information, vehicle speed information, and acceleration status from the fault identification unit. Based on the vehicle's nonlinear dynamics model, it solves the optimization problem of the maximum yaw rate under the remaining actuator execution capability, and obtains the residual maximum allowable yaw rate. Using a parametric surface fitting model or a lookup table method, it quickly retrieves the range of the vehicle's currently available maximum yaw rate. This unit outputs the vehicle residual steering capability envelope, which is a three-dimensional envelope of maximum yaw rate-vehicle speed-acceleration under specific fault conditions, providing kinematic feasibility constraints for online trajectory planning, and sends it to the online safe parking control module.

[0049] (4) Online safe parking control module: Under severe steering failure conditions, this module plans and tracks a safe parking trajectory to achieve safe parking and collision avoidance. This module receives sensor information from the sensor combination, execution fault status information from the fault identification unit, and error safety boundary, error feedback matrix, and residual steering capability from the offline fault safety boundary calculation module. In one embodiment of the present invention, the online safe parking control module further includes a safe parking target planning unit, a safe collision avoidance calculation unit, a parking trajectory optimization unit, a robust feedback control unit, and an execution command allocation unit.

[0050] Specifically, the safe parking planning unit searches for and determines the safe parking target point and the drivable area boundary, taking into account the error safety boundary, based on environmental perception information and the current status of the vehicle, and sends it to the parking trajectory optimization unit.

[0051] The safety collision avoidance calculation unit performs geometric modeling of surrounding stationary and moving obstacles, uses a convex polygon two-dimensional envelope to surround the faulty vehicle and surrounding obstacles, uses a smoothing function to process the directed distance constraints between the obstacle and the vehicle contour, generates differentiable obstacle collision avoidance constraints that take motion error into account and safety boundaries, and sends them to the parking trajectory optimization unit.

[0052] The parking trajectory optimization unit, based on an optimal control framework, integrates the vehicle's residual steering capability, error safety boundary, collision avoidance constraints, parking target, feasible region, and vehicle dynamics model into a nonlinear optimization problem. Under the premise of satisfying all constraints, this unit plans an optimal target trajectory from the current position to a safe parking point and the corresponding nominal control command, ensuring the vehicle's parking trajectory is safe and dynamically feasible, and sends the trajectory and nominal control command to the robust feedback control unit. Optionally, to ensure the current optimization is feasible and accelerate the optimization speed, a hybrid A / B control system can be used. The trajectory optimization algorithm finds a feasible trajectory as the initial value for optimization.

[0053] The robust feedback control unit receives the target trajectory and nominal control commands, calculates the deviation between the actual vehicle state and the target trajectory, and uses the error feedback matrix obtained offline to calculate robust feedback control commands in real time to offset disturbances and model errors, ensuring that the actual vehicle trajectory always stays within the error safety boundary and ensuring collision avoidance safety during the vehicle's safe parking process.

[0054] The execution command allocation unit receives actuator control commands from the robust feedback control unit, generates the final front wheel steering angle command, four-wheel braking torque command, and four-wheel drive torque command, and sends them to the actuator subsystem.

[0055] (5) The actuator subsystem includes: a four-wheel independent braking brake system, a four-wheel independent drive system, a steer-by-wire system and sensor combinations of the corresponding subsystems, and the actuator subsystem execution command generated by the actuator subsystem execution path tracking module.

[0056] Therefore, the system proposed in this application embodiment can ensure that the vehicle can make full use of its remaining differential actuation capability under severe steering failure conditions, and safely move the vehicle out of the active lane and complete the parking process while ensuring kinematic feasibility and robust collision avoidance safety.

[0057] Example 2 This invention relates to a differential steering safety parking control method for a vehicle in a severe steering failure scenario, which is implemented using the differential steering safety parking control system shown in Example 1. Figure 2 As shown, the method includes the following steps: S1 acquires vehicle dynamics status, fault information and environmental information, establishes vehicle path tracking dynamics model, and determines the remaining execution capacity and allowable execution range of each actuator based on actuator fault information.

[0058] In this embodiment of the invention, modeling is performed for vehicle path tracking dynamics that take into account differential motion:

[0059]

[0060] In the formula, This represents the vehicle's longitudinal position in the geodetic coordinate system. This represents the horizontal position in the geodetic coordinate system. For the longitudinal speed of the vehicle, The longitudinal acceleration of the vehicle; The yaw angle of the vehicle. Let yaw rate be the vehicle's angular velocity. The yaw acceleration of the vehicle; The lateral speed of the vehicle. This refers to the vehicle's lateral acceleration. Let Z be the vehicle's moment of inertia along the z-axis. For the overall vehicle weight; This refers to the front wheel steering angle. ( (representing the front and rear axles respectively) represents the lateral forces on the front and rear axles. For the longitudinal forces on the front and rear axles, For four-wheel drive / braking force command ( (These represent the front left, front right, rear left, and rear right wheels, respectively). For four-wheel drive braking torque, and , The radius of the wheel; and These are the distances from the front axle to the center of gravity and the rear axle to the center of gravity, respectively. The wheelbase of the vehicle; This refers to the actual yaw moment generated by the four-wheel differential. For modeling errors or unknown disturbances; system state Control input The above system can then be expressed as a nonlinear affine system:

[0061] in, Let be the system's autonomous function matrix. This is the system control input gain function matrix.

[0062] Furthermore, embodiments of the present invention acquire the current state of the vehicle through a combination of sensors. And calculate the information of fixed obstacles or dynamic obstacles as This represents the geometric information of the obstacle and its positional relationship over time. A fixed obstacle is assumed to be... Its position does not change over time. , This represents the number of obstacles.

[0063] Simultaneously, based on the actuator status feedback from the fault identification unit, the remaining actuator capability under vehicle execution failure conditions is calculated. The following model is performed based on the fault information:

[0064]

[0065] in, , This is a matrix of positive and negative fault information for vehicle actuators. These are the positive residual actuator capacity coefficients for the steering system and the four-wheel drive / braking system, respectively, representing the ratio of residual actuator capacity to actuator capacity under fault-free conditions; The negative residual actuator capacity coefficient of the execution system; Let be a function that generates a matrix based on its diagonal elements. Then the permissible execution range of the vehicle actuator is:

[0066] in, The execution command for the i-th executor ( These represent the front wheel steering angle command and the left front, right front, left rear, and right rear wheel braking commands, respectively. The maximum and minimum values ​​for the i-th execution subsystem under healthy conditions.

[0067] S2, based on the fault information, calculate the vehicle state error safety boundary, control input error safety boundary and corresponding state feedback matrix through the robust control invariant set.

[0068] In this embodiment of the invention, the error safety boundary and error feedback control law of the current vehicle are determined based on the fault information.

[0069] Specifically, this step, based on the modeling and estimation of model uncertainties, fault disturbances, and external disturbances under fault conditions, calculates the error safety boundary of the vehicle state through robust control invariant sets. Error safety boundary with vehicle control input :

[0070] in, System status With open-loop trajectory status of autonomous driving error; For system control input With open-loop trajectory status input for autonomous driving error; This is the upper bound of the norm of the error boundary; Let be the infinite norm of the variable.

[0071] Simultaneously, the state feedback matrix under specific fault conditions is obtained. This matrix represents the vehicle state. function.

[0072] It should be noted that the error safety boundary and error feedback control law of the current vehicle determined by the fault conditions in this step can be solved offline, and the solution results corresponding to different faults can be looked up online to reduce the computational burden.

[0073] S3. Construct a residual steering capability envelope model based on the fault information and determine the steering capability constraint range of the vehicle under fault conditions.

[0074] In step S3, the residual steering capability of the vehicle is determined based on the fault information. Considering that after steering failure, the vehicle relies on differential drive / braking to generate yaw moment, and this capability is limited by the tire friction circle and strongly coupled with longitudinal acceleration and deceleration, a residual steering capability envelope model is constructed. A cubic polynomial surface is then used to fit the envelope of the vehicle's maximum residual yaw rate under a specific fault.

[0075] Specifically, the embodiments of the present invention first construct the following optimization problem based on the differential vehicle path tracking dynamics:

[0076] In the formula, , These are the minimum and maximum values ​​of the front wheel steering angle that can be executed under fault-free conditions; , These represent the maximum braking and driving values ​​that can be performed on the four wheels under fault-free conditions, respectively. The road adhesion coefficient; The vertical load is for the four wheels; The acceleration conditions are set.

[0077] Under specific fault conditions, the speed With acceleration By iterating through the problem and solving the optimization problem above, we can obtain the following results: Figure 3The maximum yaw rate is shown at the operating point. A cubic polynomial surface fitting is performed on this surface, and the maximum vehicle yaw rate is obtained by using two polynomial surfaces and taking the minimum value:

[0078] in, This represents a surface fitted by a cubic polynomial. These are the fitting coefficients. Due to the symmetry of the vehicle model and steering, the minimum values ​​of vehicle negative acceleration and vehicle yaw rate are... Both can be obtained through symmetry, which reduces the computational difficulty.

[0079] The range of the quantitative envelope constraint for residual steering capability under vehicle fault conditions. for:

[0080] This step determines the quantization envelope constraint of the vehicle's residual steering capability based on the fault conditions. This constraint can be solved offline, and the solution results for different faults can be looked up online to reduce the computational burden.

[0081] S4. Determine safe parking areas and feasible areas based on environmental information.

[0082] In step S4, a safe parking area is determined based on environmental information.

[0083] Specifically, embodiments of the present invention, based on high-precision maps or real-time perception results, search for convex polygonal regions within the right shoulder of the road or emergency lane that meet the vehicle's geometric size requirements, defining them as terminal security sets. Define the feasible area for vehicle status based on road boundaries and vehicle status restrictions. .

[0084] S5 predicts the trajectory of obstacles, calculates the directed distance between the vehicle and the obstacle, and constructs collision avoidance constraints.

[0085] In step S5, the vehicle's trajectory is predicted, and the directed distance between the optimized trajectory and the obstacle is calculated. To handle collision avoidance constraints in the gradient optimization algorithm, this embodiment of the invention uses an approximation function to smoothly approximate the directed distance between the vehicle and the obstacle.

[0086] Specifically, the envelope of the faulty vehicle (self-driving vehicle). The directed distance to an obstacle can be calculated using the directed distance of a convex polygon. (Convex polygon) The o-th vertex To convex polygon Two clockwise vertices and The defined boundary of the e-th hyperplane distance It can be calculated as follows:

[0087]

[0088] in, Given a rotation matrix, Let be the quadratic norm of the variable.

[0089] The approximate formula is:

[0090] in, It is an adjustable parameter. and The operations of approximating the maximum and minimum values ​​can be performed respectively. For function parameters, For parameter dimensions.

[0091] The envelope of the faulty vehicle and obstacles Directed distance It can be approximated as:

[0092]

[0093]

[0094]

[0095] in, , , , ; and Convex polygons The number of vertices.

[0096] Considering the increased vehicle driving error caused by the malfunction, in order to ensure the vehicle's collision avoidance safety, the collision avoidance constraints of this application are designed as follows:

[0097] in, This is a safety margin for vehicle collisions. The Minkowski sum is a convex set. By extending the vehicle's envelope to account for driving errors, higher collision safety is ensured.

[0098] S6, based on the vehicle dynamics model, residual steering capability constraints, error safety boundaries and collision avoidance constraints, constructs a nonlinear optimization problem and solves it to obtain the nominal state trajectory and nominal control input.

[0099] In step S6, this embodiment of the application is based on a nonlinear parking optimization control problem that integrates the parking target, vehicle residual steering ability, error safety boundary, collision avoidance constraint, and vehicle dynamics model into a variable time domain. The optimization objective is to generate a nominal trajectory sequence that can quickly and safely stop the vehicle. and nominal control input sequence The optimization problem is:

[0100] in, For the system discrete time, To optimize the discrete time of the optimal system, Optimize variables for system state. For system control optimization variables, The yaw rate is the optimization variable in the system state. , The dimension is the discrete state. Weighted by parking time; The permissible range of control commands to be executed after a vehicle malfunction; For the initial optimized state of the vehicle, corresponding exist Time value, When the vehicle is parked, the corresponding exist Time value; The current vehicle status; Minkowski difference for convex sets; To optimize performance metrics, it can be defined as follows: .

[0101] Finally, the above problem is solved using a nonlinear programming solver to obtain the optimal nominal control input sequence. and vehicle status trajectory It achieves safe parking planning and open-loop control in a short time, ensuring that the planned trajectory can still achieve accurate collision avoidance planning even under severe steering failure conditions.

[0102] S7 calculates robust execution control inputs based on the state feedback matrix, according to the deviation between the actual vehicle state and the nominal trajectory.

[0103] In step S7, robust execution control input is calculated based on the deviation between the planned trajectory and the actual state of the vehicle.

[0104] Specifically, based on the feedback gain matrix calculated in step S2 Then robustly execute control input It can be calculated as follows:

[0105] This allows for the generation of real-time control commands to bring the vehicle back to its nominal trajectory, ensuring that the vehicle remains within the nominal trajectory. Hard error range Inside.

[0106] S8 controls the vehicle actuators to execute corresponding control commands based on the control input.

[0107] In step S8, the actuator executes the corresponding control command, and the four-wheel drive / braking torque... Positive values ​​are executed by the corresponding wheel's drive system, while negative values ​​are coordinated and controlled by the corresponding wheel's braking and drive systems. (Front wheel steering angle command) This is performed by the steering system.

[0108] In summary, this invention addresses the extreme condition of severe steering failure in autonomous vehicles by constructing an architecture that combines offline fault-safe boundary calculation with online safe parking control. This effectively solves the collision avoidance safety problem caused by abrupt changes in vehicle dynamics after a failure, as well as the problem of insufficient real-time computation. This invention utilizes parametric surface fitting to realize the vehicle's residual steering capability and, based on this, constructs a variable-time-domain nonlinear optimal control problem that includes safety error boundaries, smooth collision avoidance constraints, and residual steering capability constraints. Through the synergistic effect of nominal trajectory planning and robust feedback control, this invention ensures that when the vehicle loses its main steering capability, it can fully utilize differential drive / braking potential to achieve robust and accurate autonomous obstacle avoidance and safe parking while strictly satisfying dynamic constraints and physical safety boundaries.

[0109] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0110] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0111] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A differential steering safety parking control system for a vehicle steering failure scenario, characterized in that, It includes a sensor array, a fault identification unit, an offline fault-safe boundary calculation module, an online safe parking control module, and an actuator subsystem, wherein: The sensor array is responsible for collecting vehicle dynamics data in real time, fusing environmental data using visual and radar algorithms, and finally sending the processed vehicle motion state and environmental perception information to the fault identification unit and online safe parking control module in real time. The fault identification unit is used to accurately identify the specific failure location of the steering system, drive system or braking system by analyzing the response residual, current feedback or hydraulic state and vehicle dynamic state of the actuator fed back by the sensor combination, and further calculate the remaining efficiency ratio and asymmetric fault characteristic parameters of the failed actuator, thereby generating a vehicle actuator fault state feedback signal containing the fault location and positive and negative remaining efficiency coefficients, and sending it to the offline fault safety boundary calculation module and the online safe parking control module in real time. The offline fault safety boundary calculation module is used to receive fault feedback information from the fault identification unit and provide the online safe parking control module with error safety boundary, error feedback control matrix and residual steering capability signal; The online safe parking control module is used to receive sensing information from the sensor combination, execution fault status information from the fault identification unit, and error safety boundary, error feedback matrix, and residual steering capability from the offline fault safety boundary calculation module. Under severe steering failure conditions, it plans and tracks a safe parking trajectory to achieve safe parking and collision avoidance. The actuator subsystem includes a four-wheel independent braking brake system, a four-wheel independent drive system, a steer-by-wire system, and sensor combinations of the corresponding subsystems, used to execute the generated actuator subsystem execution commands.

2. The system according to claim 1, characterized in that, The offline fault-safe boundary calculation module includes: The error safety boundary calculation unit, based on the pre-calibrated model error distribution, calculates the convergence range of the vehicle's state error under closed-loop control for the current fault mode, and generates the error safety boundary and the corresponding error feedback matrix, which are used to quantify the maximum trajectory deviation of the vehicle under faults and disturbances. The vehicle residual steering capability calculation unit receives fault information, vehicle speed information, and acceleration status from the fault identification unit. Based on the vehicle's nonlinear dynamics model, it solves the optimization problem of the maximum yaw rate under the remaining actuator execution capability, obtains the residual maximum allowable yaw rate, and quickly retrieves the current available maximum yaw rate range of the vehicle using a parametric surface fitting model or a lookup table method.

3. The method according to claim 2, characterized in that, The online safe parking control module includes a safe parking target planning unit, a safe collision avoidance calculation unit, a parking trajectory optimization unit, a robust feedback control unit, and an execution command allocation unit, wherein: The safe parking planning unit searches for and determines the safe parking target point and the drivable area boundary, taking into account the error safety boundary, based on the environmental perception information and the current state of the vehicle, and sends them to the parking trajectory optimization unit. The safety collision avoidance calculation unit is used to perform geometric modeling of surrounding stationary and moving obstacles, use a convex polygon two-dimensional envelope to surround the faulty vehicle and surrounding obstacles, and use a smoothing function to process the directed distance constraint between the obstacle and the vehicle contour to generate a differentiable obstacle collision avoidance constraint that considers the safety boundary of motion error, and send it to the parking trajectory optimization unit. The parking trajectory optimization unit is used to integrate the vehicle's residual steering ability, error safety boundary, collision avoidance constraints, parking target, feasible area and vehicle dynamics model into a nonlinear optimization problem based on the optimal control framework. Under the premise of satisfying all constraints, it plans the optimal target trajectory from the current position to the safe parking point and the corresponding nominal control command, ensuring that the vehicle parking trajectory is safe and dynamically feasible, and sends the trajectory and nominal control command to the robust feedback control unit. The robust feedback control unit receives the target trajectory and nominal control commands and calculates the deviation between the actual vehicle state and the target trajectory. Using the error feedback matrix obtained offline, it calculates robust feedback control commands in real time to offset disturbances and model errors, ensuring that the actual vehicle trajectory always remains within the error safety boundary. The execution command allocation unit is used to receive actuator control commands from the robust feedback control unit, generate the final front wheel steering angle command, four-wheel braking torque command, and four-wheel drive torque command, and send them to the actuator subsystem.

4. A differential steering safety stopping control method for a vehicle in a severe steering failure scenario, characterized in that, Implemented by the differential steering safety parking control system according to any one of claims 1-3, comprising: Acquire vehicle dynamics status, fault information and environmental information, establish a vehicle path tracking dynamics model, and determine the remaining execution capacity and allowable execution range of each actuator based on actuator fault information; Based on the fault information, the vehicle state error safety boundary, the control input error safety boundary, and the corresponding state feedback matrix are calculated using the robust control invariant set. Based on the fault information, a residual steering capability envelope model is constructed, and the steering capability constraint range of the vehicle under fault conditions is determined. Determine safe parking areas and feasible areas based on vehicle status based on environmental information; Predict the trajectory of obstacles, calculate the directed distance between the vehicle and the obstacle, and construct collision avoidance constraints; Based on the vehicle dynamics model, residual steering capability constraints, error safety boundary and collision avoidance constraints, a nonlinear optimization problem is constructed and solved to obtain the nominal state trajectory and nominal control input. Based on the deviation between the actual vehicle state and the nominal trajectory, robust execution control input is calculated using the state feedback matrix. The vehicle actuators are controlled to execute corresponding control commands based on the control input.

5. The method according to claim 4, characterized in that, Acquire vehicle dynamics status, fault information, and environmental information; establish a vehicle path tracking dynamics model; and determine the remaining execution capacity and permissible execution range of each actuator based on actuator fault information, including: Modeling is performed for vehicle path tracking dynamics that consider differential motion: In the formula, This represents the vehicle's longitudinal position in the geodetic coordinate system. This represents the horizontal position in the geodetic coordinate system. For the longitudinal speed of the vehicle, The longitudinal acceleration of the vehicle; The yaw angle of the vehicle. Let yaw rate be the vehicle's angular velocity. The yaw acceleration of the vehicle; The lateral speed of the vehicle. This refers to the vehicle's lateral acceleration. Let Z be the vehicle's moment of inertia along the z-axis. For the overall vehicle weight; This refers to the front wheel steering angle. ( (representing the front and rear axles respectively) represents the lateral forces on the front and rear axles. For the longitudinal forces on the front and rear axles, For four-wheel drive / braking force command ( (These represent the front left, front right, rear left, and rear right wheels, respectively). For four-wheel drive braking torque, and , The radius of the wheel; and These are the distances from the front axle to the center of gravity and the rear axle to the center of gravity, respectively. The wheelbase of the vehicle; This refers to the actual yaw moment generated by the four-wheel differential. For modeling errors or unknown disturbances; system state Control input ; The above system can be expressed as a nonlinear affine system: in, Let be the system's autonomous function matrix. The system control input gain function matrix; The current status of the vehicle is obtained through a combination of sensors. And calculate the information of fixed obstacles or dynamic obstacles as This represents the geometric information of the obstacle and its positional relationship over time. A fixed obstacle is assumed to be... Its position does not change over time. , Represents the number of obstacles; Based on the actuator status feedback from the fault identification unit, the remaining actuator capability under vehicle execution failure conditions is calculated, and the following model is performed based on the fault information: in, , This is a matrix of positive and negative fault information for vehicle actuators. These are the positive residual actuator capacity coefficients for the steering system and the four-wheel drive / braking system, respectively, representing the ratio of residual actuator capacity to actuator capacity under fault-free conditions; The negative residual actuator capacity coefficient of the execution system; Let be a function that generates a matrix based on its diagonal elements. Then the permissible execution range of the vehicle actuator is: in, The execution command for the i-th executor ( These represent the front wheel steering angle command and the left front, right front, left rear, and right rear wheel braking commands, respectively. The maximum and minimum values ​​for the i-th execution subsystem under healthy conditions.

6. The method according to claim 5, characterized in that, Based on the fault information, the vehicle state error safety boundary, control input error safety boundary, and corresponding state feedback matrix are calculated using robust control invariant sets, including: Based on the modeling and estimation of model uncertainties, fault disturbances, and external disturbances under fault conditions, the error safety boundary of the vehicle state is obtained through robust control invariant set calculation. Error safety boundary with vehicle control input : in, System status With open-loop trajectory status of autonomous driving error; For system control input With open-loop trajectory status input for autonomous driving error; This is the upper bound of the norm of the error boundary; The infinite norm of the variable; Simultaneously, the state feedback matrix under specific fault conditions is obtained. This matrix represents the vehicle state. function.

7. The method according to claim 6, characterized in that, Based on the fault information, a residual steering capability envelope model is constructed, and the steering capability constraint range of the vehicle under fault conditions is determined, including: The following optimization problem is constructed by combining differential vehicle path tracking dynamics: In the formula, , These are the minimum and maximum values ​​of the front wheel steering angle that can be executed under fault-free conditions; , These represent the maximum braking and driving values ​​that can be performed on the four wheels under fault-free conditions, respectively. The road adhesion coefficient; The vertical load is for the four wheels; The acceleration condition is set. Under specific fault conditions, the speed With acceleration The above optimization problem is solved by iterating through the data to obtain the operating point with the maximum yaw rate. A cubic polynomial surface fitting is then performed on the resulting surface. The maximum yaw rate of the vehicle is obtained by using two polynomial surfaces and taking the minimum value. in, This represents a surface fitted by a cubic polynomial. These are the fitting coefficients; Based on the symmetry and steering symmetry of the vehicle model, the minimum values ​​of vehicle negative acceleration and vehicle yaw rate are obtained. ; Calculate the quantitative envelope constraint range of residual steering capability under vehicle fault conditions. for: The solution is complete.

8. The method according to claim 7, characterized in that, Based on environmental information, safe parking areas and areas where vehicle status is feasible are determined, including: Based on high-precision maps or real-time sensing results, a convex polygon region that meets the vehicle's geometric dimensions is searched within the right shoulder of the road or emergency lane, and defined as the terminal safety set. ; Define the feasible area for vehicle status based on road boundaries and vehicle status restrictions. .

9. The method according to claim 8, characterized in that, Predict the trajectory of obstacles, calculate the directed distance between the vehicle and the obstacle, and construct collision avoidance constraints, including: Calculate the faulty vehicle using the directed distance of a convex polygon. The directed distance to the obstacle, where the convex polygon The o-th vertex To convex polygon Two clockwise vertices and The defined boundary of the e-th hyperplane distance The calculation is as follows: in, Given a rotation matrix, Let be the quadratic norm of the variable; The approximate formula is: in, It is an adjustable parameter. and The operations of approximating the maximum and minimum values ​​can be performed respectively. For function parameters, For parameter dimensions; The envelope of the faulty vehicle and obstacles Directed distance It can be approximated as: in, , , , ; and Convex polygons The number of vertices; The collision avoidance constraints for the vehicle are designed as follows: in, This is a safety margin for vehicle collisions. Minkowski sum for convex sets.

10. The method according to claim 9, characterized in that, Based on the vehicle dynamics model, residual steering capability constraints, error safety boundaries, and collision avoidance constraints, a nonlinear optimization problem is constructed and solved to obtain the nominal state trajectory and nominal control input, including: This problem integrates the parking objective, residual vehicle steering capability, error safety boundary, collision avoidance constraints, and vehicle dynamics model into a nonlinear parking optimization control problem in the variable time domain. The optimization objective is to generate a nominal trajectory sequence that enables a fast and safe stop. and nominal control input sequence The optimization problem is: in, For the system discrete time, To optimize the discrete time of the optimal system, Optimize variables for system state. For system control optimization variables, The yaw rate is the optimization variable in the system state. , The dimension is the discrete state. Weighted by parking time; The permissible range of control commands to be executed after a vehicle malfunction; For the initial optimized state of the vehicle, corresponding exist Time value, When the vehicle is parked, the corresponding exist Time value; The current vehicle status; Minkowski difference for convex sets; To optimize performance metrics, it can be defined as follows: ; The above problem is solved using a nonlinear programming solver to obtain the optimal nominal control input sequence. and vehicle status trajectory .

Images