Multi-domain cooperative stability control system and method for intelligent electrically driven heavy-duty vehicle

Abstract

The application provides a multi-domain cooperative stability control system and method for an intelligent electrically-driven heavy-load vehicle, which builds a control architecture composed of a smart driving domain, a chassis domain and a VCU, predicts a front road condition by the smart driving domain, plans a track based on the front road condition, analyzes the stability of the vehicle according to the state of the vehicle, determines a safety boundary, reallocates the torque of four wheels of the vehicle by the chassis domain, re-plans the track of the vehicle according to the real-time state of the vehicle, tracks the planned track by using an AMPC method, and finally controls and executes corresponding torque and steering by the VCU, so that the multi-domain cooperative stability control system and safety boundary calculation suitable for the intelligent electrically-driven heavy-load vehicle such as a mine dump truck are realized, and the stability of the vehicle is effectively improved.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of automatic driving of electric vehicles, and particularly relates to a multi-domain cooperative stability control system and method for intelligent electrically-driven heavy-load vehicles. BACKGROUND

[0002] As a typical automatic driving application scenario, open-pit mines have more complex working conditions than small passenger cars driving on structured urban roads. Intelligent mine dump trucks and other intelligent electrically-driven heavy-load vehicles work on the gravel roads of mines for a long time, and their operating conditions have the characteristics of relatively fixed route, large slope ratio, large single load of vehicles, high energy consumption, and high failure rate in continuous operation scenarios of vehicles. Therefore, ensuring the efficiency, safety and reliability of the vehicle during operation is the primary goal in the design of the automatic driving control strategy of such vehicles. To ensure the safety of the vehicle, the stability control of intelligent electrically-driven heavy-load vehicles is one of the key research objects of the control system, and one of the prerequisites for solving the stability control problem is to determine the safety boundary of the vehicle. However, the existing stability control and safety boundary calculation methods still have some shortcomings to be improved. SUMMARY

[0003] Therefore, in view of the technical problems existing in the prior art, the present application provides a multi-domain cooperative stability control system for intelligent electrically-driven heavy-load vehicles, which is composed of various sensors, chassis domain controllers and automatic driving domain controllers.

[0004] Among them, the various sensors are used to collect various sensing signals during the driving of the vehicle.

[0005] The automatic driving domain controller is specifically composed of a road adhesion coefficient estimation module, a road condition prediction module, a yaw angular velocity module, a safety boundary analysis module and a driving state prediction module. The road adhesion coefficient estimation module is used to comprehensively calculate based on the acceleration signal, the front wheel angle signal, the wheel angular velocity signal and the wheel torque signal, and output the road adhesion coefficient estimation value. The road condition prediction module is used to predict the front road adhesion coefficient by image recognition of the camera signal, and to perform spatio-temporal fusion with the road adhesion coefficient estimation value obtained by the road adhesion coefficient estimation module to obtain a reliable road adhesion coefficient. The yaw angular velocity module is used to calculate and output the vehicle yaw angular velocity based on the lateral acceleration and the vehicle speed. The trajectory planning module plans the trajectory according to the real-time pose and real-time road condition information of the vehicle. The safety boundary analysis module is used to fuse the road adhesion coefficient, the vehicle yaw angular velocity and the sensing signals obtained by the various sensors to calculate the vehicle dynamics boundary including the yaw angular velocity and the mass center side slip angle, so as to analyze the current stability state of the vehicle. The driving state prediction module is used to predict the vehicle stability state in the subsequent period of time according to the current stability state and other vehicle information.

[0006] The chassis domain controller is specifically composed of a steering control module, a speed control module, a driving trajectory re-planning module, and a trajectory tracking module; wherein the steering control module and the speed control module are used to control the steering and speed of the vehicle according to the vehicle stability state predicted by the driving state prediction module, and output the steering and speed boundary values at which the vehicle can stably drive; the driving trajectory re-planning module is used to determine whether trajectory re-planning is needed according to the road adhesion coefficient provided by the road condition prediction module, the vehicle stability state predicted by the driving state prediction module, and the steering and speed boundary values; and the trajectory tracking module is used to execute the AMPC algorithm to track and control the trajectory, and output the torque signal and the steering signal of each wheel of the vehicle.

[0007] Further, the reliable road adhesion coefficient is obtained through the following steps:

[0008] The road adhesion coefficient estimation module first establishes a four-wheel seven-degree-of-freedom model for the vehicle, and the longitudinal forces borne by the four wheels are expressed as:

[0009]

[0010] wherein m is the mass of the vehicle, g is the acceleration of gravity, a x and a y are the longitudinal and lateral accelerations of the vehicle respectively, t f and t r are the wheelbase of the front axle and the rear axle respectively, h g represents the height of the vehicle's center of mass, r is the yaw rate, b is the length of the rear axle from the center of mass of the vehicle, and L represents the wheelbase of the vehicle.

[0011] The HSRI model is used to calculate the longitudinal force F x of the tire according to the force borne by the tire:

[0012]

[0013] wherein,

[0014]

[0015] μ=μ0(1-A s V s )

[0016]

[0017] C s , C α represent the longitudinal slip stiffness and the cornering stiffness of the tire respectively, s x , s y represent the longitudinal slip ratio and the lateral slip ratio of the tire respectively, and F zis the longitudinal force on the tire, λ is a coefficient related to the working condition, λ = 0 represents the working condition of wheel lock, μ is the road adhesion coefficient, μ 0 represents the peak value, A s is the speed influence factor, V is the tire driving speed, and a is the tire side slip angle;

[0018] An algorithm for estimating the road adhesion coefficient based on extended Kalman filtering is established, and the state variables are defined where r is the yaw rate, is the roll angle, is the roll rate, τ is the yaw moment, and a is the tire side slip angle ij is the tire side slip angle, s ij is the tire longitudinal slip ratio, F z_ij is the vertical load of the vehicle, and x is the parameter variable p = [μ fl , μ fr , μ rl , μ rr ], μ is the tire road adhesion coefficient, and the subscripts fl, fr, rl, and rr represent the left front, right front, left rear, and right rear four wheels, respectively, u(t) = [δ ij , w fl , w fr , w rl , w rr ] is the control input, δ is the wheel angle, and w is the tire rolling angular velocity, and the test output is a x , a y , represent the longitudinal acceleration, lateral acceleration, and yaw angular acceleration, respectively, and the road adhesion coefficient estimate is calculated therefrom;

[0019] The road condition prediction module specifically fuses the road adhesion coefficient estimate with the visual sensor and obtains a reliable road adhesion coefficient in front of the vehicle through a deep learning method.

[0020] Further, the safety boundary analysis module obtains the vehicle dynamics boundary including the yaw rate and the mass center side slip angle through the following steps:

[0021] First, the vehicle model is simplified to a single-track bicycle template and subjected to force analysis, and the following state space equation of the vehicle is obtained:

[0022]

[0023] In the formula, F yf , F yr are the front wheel lateral force and the rear wheel lateral force, β is the mass center side slip angle, r is the yaw rate, and the superscript · represents the derivative of the corresponding parameter;

[0024] The stability of the vehicle is characterized by the centroid side slip angle- yaw rate phase plane, i.e. a parallelogram stability region is defined on the β-r phase plane, with the upper and lower boundaries determined by the maximum yaw rate under the current tire-road friction, and the left and right boundaries come from the constraint that the rear wheel slip angle is kept in the unsaturated region; considering that the lateral force of the rear wheel is saturated before the front wheel in the stability design of the vehicle, the maximum yaw rate r max :

[0025]

[0026] Similarly, the limit of the centroid side slip angle is obtained according to the limit of the rear wheel side slip:

[0027] -∝ max,r ≤≤∝ r ≤∝ max,r

[0028]

[0029] In the formula, μ is the adhesion coefficient of the driving road, ∝ is the tire side slip angle of the vehicle, g is the gravitational acceleration, and a and b are the lengths of the front and rear axles of the vehicle from the vehicle centroid.

[0030] Further, the trajectory tracking module specifically tracks the trajectory by the following AMPC method:

[0031] First, the vehicle dynamics model is simplified to the following form:

[0032]

[0033] A cost function is designed, and the minimum of the cost function is taken as the optimization objective to obtain the control amount deviation of the control step; the specific form of the cost function is as follows:

[0034] J k =J zk +J uk +J duk +J εk

[0035] Among them, J zk , J uk , J duk are the cumulative cost functions of the tracking error, the cumulative cost function of the control amount, and the change cumulative cost function of the optimized control amount in the prediction time window, which are used to ensure the tracking of the vehicle to the target trajectory, and J εk is the cost function of the relaxation variable, and its specific form is:

[0036]

[0037] wherein x(k+h|k) is the vehicle state at k+h| time, u(k+h) is the vehicle action at k+h| time, Q(k+h||k), R(k), E(k) are all the weight matrices corresponding to k+h time, S z is a diagonal matrix of normalization coefficients, s d , s du are normalization coefficients;

[0038] Then the multi-objective optimization problem based on AMPC is represented as:

[0039]

[0040]

[0041] wherein u ass,max is the maximum allowed correction, u max is the limit value of the actuator.

[0042] Correspondingly, the application also provides a multi-domain collaborative stability control method for an intelligent electrically driven heavy-duty vehicle, which is executed by using the above system, and specifically includes the following steps:

[0043] S1, the road adhesion coefficient of the road on which the vehicle currently locates is estimated in real time by a road adhesion coefficient estimation module by using the vehicle sensing signals provided by various sensors, and the yaw rate is calculated by a yaw rate module; the reliable road adhesion coefficient of the front road section is obtained by fusing the road adhesion coefficient estimation value and the visual sensor by a road condition prediction module;

[0044] S2, the sensing signals and the data obtained in step S1 are integrated by a safety boundary analysis module, and the stability of the vehicle is represented by a centric side slip angle-yaw rate phase plane; the vehicle yaw rate and the centric side slip angle constraints under the maximum performance of the vehicle rear wheel tire are set as the stability constraints for the stability envelope of the vehicle posture;

[0045] S3, the safety boundary of the vehicle is established by a steering control module and a speed control module based on the stability constraints obtained in S2, the real-time state point of the vehicle is mapped, the stability state in which the vehicle locates and the stability risk at the future time are judged, and the steering and speed boundary values under which the vehicle can stably travel are output;

[0046] S4, the driving trajectory of the vehicle is re-planned and judged by a driving trajectory re-planning module according to the current state point of the vehicle, the safety boundary and the road adhesion coefficient estimation of the front road;

[0047] S5, the driving trajectory of the vehicle is tracked by a trajectory tracking module based on the AMPC mode, and the driving, braking and steering systems of the vehicle are controlled.

[0048] The multi-domain cooperative stability control system and method of the intelligent electrically driven heavy-load vehicle provided by the application build a control architecture composed of a chassis domain, an intelligent driving domain and a VCU, the chassis domain predicts the road conditions ahead, analyzes the stability of the vehicle state and determines the safety boundary in combination with the vehicle state, the intelligent driving domain reallocates the torque of the four wheels of the vehicle and re-plans the driving trajectory of the vehicle, and the AMPC method is used to track the vehicle dynamics of the newly planned trajectory, and finally the corresponding torque and steering are controlled by the VCU to execute, so as to realize the multi-domain cooperative stability control system and safety boundary calculation suitable for the intelligent electrically driven heavy-load vehicle such as a mine dump truck, and effectively improve the stability of the vehicle. BRIEF DESCRIPTION OF DRAWINGS

[0049] Figure 1 The frame diagram of the system provided by the application is shown in the figure;

[0050] Figure 2 The flowchart of the method provided by the application is shown in the figure;

[0051] Figure 3 The vehicle yaw rate and mass center side slip angle plane diagram and parallelogram safety boundary diagram are shown in the figure;

[0052] Figure 4 The optional control frame diagram of the driving trajectory tracking based on AMPC is shown in the figure. DETAILED DESCRIPTION

[0053] The technical solutions of the application will be described clearly and completely in combination with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the application, rather than all the embodiments. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the protection scope of the application.

[0054] The multi-domain cooperative stability control system of the intelligent electrically driven heavy-load vehicle provided by the application is composed of various sensors, a chassis domain controller and an automatic driving domain controller, as shown in the figure. Figure 1

[0055] Among them, the various sensors are used to collect various sensing signals in the vehicle driving process.

[0056] ​The automatic driving domain controller is specifically composed of a road adhesion coefficient estimation module, a road condition prediction module, a yaw angular velocity module, a trajectory planning module, a safety boundary analysis module, and a driving state prediction module; the road adhesion coefficient estimation module is used for comprehensive calculation based on acceleration signals, front wheel angle signals, wheel angular velocity signals and wheel torque signals, and outputs a road adhesion coefficient estimation value; the road condition prediction module is used for image recognition on camera signals to predict a front road adhesion coefficient, and performs space-time fusion on the road adhesion coefficient estimation value obtained by the road adhesion coefficient estimation module to obtain a reliable road adhesion coefficient; the yaw angular velocity module is used for calculating and outputting a vehicle yaw angular velocity based on lateral acceleration and vehicle speed; the trajectory planning module performs trajectory planning according to real-time vehicle pose and real-time road condition information; the safety boundary analysis module is used for fusing the road adhesion coefficient, the vehicle yaw angular velocity and sensing signals obtained by various sensors, and calculating vehicle dynamics boundaries including a yaw angular velocity and a mass center side slip angle to analyze a current stability state of the vehicle; the driving state prediction module is used for predicting a vehicle stability state in a subsequent period of time according to the current stability state and other ego vehicle information.

[0057] The chassis domain controller is specifically composed of a steering control module, a speed control module, a driving trajectory re-planning module and a trajectory tracking module; the steering control module and the speed control module are used for controlling steering and speed of the vehicle according to the vehicle stability state predicted by the driving state prediction module, and output steering and speed boundary values at which the vehicle can stably drive; the driving trajectory re-planning module is used for judging whether trajectory re-planning is needed according to the road adhesion coefficient provided by the road condition prediction module, the vehicle stability state predicted by the driving state prediction module, the steering and speed boundary values; the trajectory tracking module is used for performing AMPC algorithm to track and control the trajectory, and output torque signals and steering signals of each wheel of the vehicle.

[0058] In the preferred embodiment of the present application, the reliable road adhesion coefficient is obtained by the following steps:

[0059] The road adhesion coefficient estimation module first establishes a four-wheel seven-degree-of-freedom model for the vehicle, and represents the longitudinal force borne by the four wheels as:

[0060]

[0061] wherein m is the mass of the vehicle, g is the acceleration of gravity, a x and a y are the longitudinal and lateral accelerations of the vehicle respectively, t f and t r are the wheelbase of the front axle and the rear axle respectively, h gis the height of the vehicle center of mass, r is the yaw rate, b is the length of the rear axle from the vehicle center of mass, and L is the wheelbase of the vehicle;

[0062] The longitudinal force F of the tire is calculated according to the force of the tire by using the HSRI model x :

[0063]

[0064] wherein,

[0065]

[0066] μ = μ0(1 - A s V s )

[0067]

[0068] C s , C α respectively represent the longitudinal slip stiffness and the cornering stiffness of the tire, s x , s y respectively represent the longitudinal slip ratio and the lateral slip ratio of the tire, F z is the longitudinal force of the tire, λ is a coefficient related to the working condition, λ = 0 represents the working condition of wheel lock, μ is the road adhesion coefficient, μ0 represents the peak value, A s is a speed influence factor, V is the driving speed of the tire, and α is the tire cornering angle.

[0069] A road adhesion coefficient estimation algorithm based on extended Kalman filtering is established, and state variables are defined wherein r is the yaw rate, is the roll angle, is the roll rate, τ is the yaw moment, and α ij is the tire cornering angle, ij is the longitudinal slip ratio of the tire, and F z_ij is the vertical load of the vehicle, and the parameter variable is x p = [μ fl , μ fr , μ rl , μ rr ], μ is the tire road adhesion coefficient, the subscripts fl, fr, rl, and rr respectively represent the left front, right front, left rear, and right rear four wheels, the control input is u(t) = [δ ij , w fl , w rr , w rl , w rr ], δ is the wheel angle, w is the tire rolling angular velocity, and the test output is a x , ay , These represent longitudinal acceleration, lateral acceleration, and yaw rate acceleration, respectively, from which the estimated value of the road adhesion coefficient is calculated.

[0070] The road condition prediction module specifically fuses the road surface adhesion coefficient estimate with the visual sensor and obtains a reliable road surface adhesion coefficient in front of the vehicle through deep learning methods.

[0071] In a preferred embodiment of the present invention, the safety boundary analysis module obtains the vehicle dynamics boundary, including yaw rate and center of gravity sideslip angle, through the following steps:

[0072] First, the vehicle model is simplified to a monorail bicycle template and subjected to force analysis, resulting in the following state-space equations for the vehicle:

[0073]

[0074] In the formula, F yf F yr Let β be the lateral force of the front wheel and the lateral force of the rear wheel, β be the sideslip angle of the center of mass, r be the yaw rate, and the superscript · indicate the derivative of the corresponding parameter;

[0075] Through such Figure 3 The centroid sideslip angle-yaw rate phase plane diagram is shown to characterize vehicle stability. Specifically, a parallelogram-shaped stability region is defined on the β-r phase plane, with its upper and lower boundaries determined by the maximum yaw rate under the current tire-road friction, and its left and right boundaries constrained by keeping the rear wheel slip angle within the unsaturated region. Considering that the lateral force on the rear wheels saturates before that on the front wheels in the vehicle's stability design, the maximum yaw rate r is thus derived. max :

[0076]

[0077] Similarly, based on the existence of a limit to rear wheel sideslip, the constraint on the center of gravity sideslip angle is obtained:

[0078] -∝ max,r ≤∝ r ≤∝ max,r

[0079]

[0080] In the formula, μ is the adhesion coefficient of the road surface, ∝ is the tire slip angle of the vehicle, g is the gravitational acceleration, and a and b are the distances from the front and rear axles of the vehicle to the center of gravity of the vehicle.

[0081] In a preferred embodiment of the present invention, the trajectory tracking module specifically performs trajectory tracking using the following AMPC method:

[0082] Firstly, the vehicle dynamics model is simplified to the following form:

[0083]

[0084] In the formula, Cc is the tire cornering stiffness.

[0085] A cost function is designed, and the control amount deviation of the control step is obtained by taking the minimum of the cost function as the optimization goal; the specific form of the cost function is as follows:

[0086] J k =J zk +J uk +J duk +J εk

[0087] Wherein, J zk , J uk , J duk are the cumulative cost functions of the tracking error, the control amount and the change of the optimized control amount in the prediction time window, respectively, which are used to ensure the tracking of the vehicle to the target trajectory, and J εk is the cost function of the relaxation variable, and the specific form is as follows:

[0088]

[0089] Wherein, x(k+h|k) is the vehicle state at k+h|, u(k+h) is the vehicle execution action at k+h|, Q(k+h|k), R(k), E(k) are all weight matrices corresponding to k+h, and S z is a diagonal matrix of normalization coefficients, s u , s du are normalization coefficients.

[0090] Then, the multi-objective optimization problem based on AMPC is represented as:

[0091]

[0092] Wherein, u ass,mxa is the maximum allowed correction amount, and u max is the limit value of the actuator.

[0093] Correspondingly, the application also provides a multi-domain collaborative stability control method for an intelligent electrically driven heavy load vehicle, which is executed by using the above system, as shown in the following formula: Figure 2 Specifically, the method comprises the following steps:

[0094] S1, the road adhesion coefficient of the road where the vehicle is currently located is estimated in real time by a road adhesion coefficient estimation module using the vehicle sensing signals provided by various sensors, and the yaw rate is calculated by a yaw rate module; the reliable road adhesion coefficient of the front road section is obtained by fusing the estimated value of the road adhesion coefficient and the visual sensor by a road condition prediction module;

[0095] S2, the safety boundary analysis module integrates the sensing signals and the data obtained in step S1, and represents the stability of the vehicle by a centric side slip angle-yaw rate phase plane; the vehicle yaw rate and centric side slip angle constraints under the maximum performance of the rear wheel tire of the vehicle are set as the stability constraints for the stability envelope of the vehicle posture;

[0096] S3, the steering control module and the speed control module establish the safety boundary of the vehicle based on the stability constraints obtained in S2, map the real-time state position point of the vehicle, judge the stability state of the vehicle and the stability risk at the future time, and output the steering and speed boundary values under which the vehicle can be stably driven;

[0097] S4, the driving trajectory re-planning module re-plans the driving trajectory of the vehicle according to the current state position point of the vehicle, the safety boundary, and the estimated road adhesion coefficient of the front road;

[0098] S5, the trajectory tracking module tracks the driving trajectory of the vehicle based on the AMPC method, controls the driving, braking and steering systems of the vehicle, and optionally predicts the tracking process as shown in Figure 4

[0099] It should be understood that the size of the serial number of each step in the embodiments of the present application does not mean the order of execution, and the execution order of each process should be determined according to its function and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0100] Although the embodiments of the present application have been shown and described, it can be understood by those of ordinary skill in the art that various changes, modifications, replacements and variations can be made to these embodiments without departing from the principles and spirits of the present application, and the scope of the present application is defined by the appended claims and their equivalents.​

Claims

1. A multi-domain collaborative stability control system for intelligent electrically driven heavy-duty vehicles, characterized in that: The plurality of sensors, the chassis domain controller, the automatic driving domain controller are composed of; The plurality of sensors are used for collecting various sensing signals in the vehicle driving process; The automatic driving domain controller is composed of a road adhesion coefficient estimation module, a road condition prediction module, a yaw angular velocity module, a trajectory planning module, a safety boundary analysis module, and a driving state prediction module; The road adhesion coefficient estimation module is used for comprehensive calculation based on acceleration signals, front wheel angle signals, wheel angular velocity signals and wheel torque signals, and outputs the road adhesion coefficient estimation value; The road condition prediction module is used for image recognition on the camera signal to predict the road adhesion coefficient in front, and is spatiotemporally fused with the road adhesion coefficient estimation value obtained by the road adhesion coefficient estimation module to obtain a reliable road adhesion coefficient; The yaw angular velocity module is used for calculating and outputting the vehicle yaw angular velocity based on the lateral acceleration and the vehicle speed; The trajectory planning module plans the trajectory according to the real-time pose of the vehicle and the real-time road condition information; The safety boundary analysis module is used for fusing the road adhesion coefficient, the vehicle yaw angular velocity and the sensing signals obtained by the plurality of sensors, and calculating the vehicle dynamics boundary including the yaw angular velocity and the centroid side slip angle to analyze the current stability state of the vehicle; The driving state prediction module is used for predicting the vehicle stability state in the subsequent period of time according to the current stability state and other ego vehicle information; The chassis domain controller is composed of a steering control module, a speed control module, a driving trajectory re-planning module and a trajectory tracking module; The steering control module and the speed control module are used for controlling the steering and the speed of the vehicle according to the vehicle stability state predicted by the driving state prediction module, and output the steering and speed boundary values at which the vehicle can stably drive; The driving trajectory re-planning module is used for judging whether the trajectory needs to be re-planned according to the road adhesion coefficient provided by the road condition prediction module, the vehicle stability state predicted by the driving state prediction module, the steering and speed boundary values; The trajectory tracking module is used for executing the AMPC algorithm to track and control the trajectory, and output the torque signal and the steering signal of each wheel of the vehicle.

2. The system of claim 1, wherein: The reliable road adhesion coefficient is obtained by the following steps: The road adhesion coefficient estimation module first establishes a four-wheel seven-degree-of-freedom model for the vehicle, and represents the longitudinal force of the four wheels considering the vertical axis load transfer as follows: where m is the mass of the vehicle, g is the acceleration due to gravity, a x and a y are the longitudinal and lateral accelerations of the vehicle, respectively, t f and t r are the track of the front and rear axles, respectively, h g denotes the height of the vehicle's center of mass, r is the yaw rate, b is the length of the rear axle from the vehicle's center of mass, and L denotes the wheelbase of the vehicle. And use the HSRI model according to the stress of the tire to calculate the longitudinal force F of the tire x : The road condition prediction module specifically fuses the road adhesion coefficient estimation value with the visual sensor, and obtains the reliable road adhesion coefficient in front of the vehicle through a deep learning method. C s , C α denote the longitudinal and lateral slip stiffness of the tire, respectively x , s y denote the longitudinal and lateral slip ratio of the tire, respectively z is the longitudinal force acting on the tire, λ is a condition-dependent coefficient, λ = 0 denotes a wheel lock condition, μ is the road adhesion coefficient, μ0 denotes its peak value, A s is a speed influence factor, V is the tire running speed, and α is the tire slip angle; Establish a road surface adhesion coefficient estimation algorithm based on extended Kalman filtering, and define the state variables. Where r is the yaw rate, For the roll angle, For the roll angular velocity, τ is the yaw moment, and a is the yaw moment. ij Tire slip angle, s ij Tire longitudinal slip ratio, F z_ij The vertical load of the vehicle is x. p =[μ fl μ fr μ rl μ rr μ is the tire-road adhesion coefficient, and the subscripts fl, fr, rl, and rr represent the four wheels: left front, right front, left rear, and right rear, respectively. The control input is u(t) = [δ] ij w fl w fr W rl w rr ], where δ is the wheel rotation angle and w is the tire rolling angular velocity, and the test output is a x a y , These represent longitudinal acceleration, lateral acceleration, and yaw rate acceleration, respectively, from which the estimated value of the road adhesion coefficient is calculated. The safety boundary analysis module obtains the vehicle dynamics boundary including the yaw angular velocity and the centroid side slip angle by the following steps:

3. The system of claim 2, wherein: First, the vehicle model is simplified into a single-track bicycle template and subjected to force analysis to obtain the following state space equation of the vehicle: Similarly, according to the limit of the rear wheel side slip, the limit of the centroid side slip angle is obtained as follows: where F yf , F yr are the front and rear lateral forces, β is the center of mass side slip angle, r is the yaw rate, and the superscript · denotes the derivative of the corresponding parameter; The stability of the vehicle is characterized by the centroid side slip angle - yaw rate phase plane, i.e. a parallelogram stability region is defined on the β-r phase plane, with the upper and lower boundaries determined by the maximum yaw rate under the current tire-road friction, and the left and right boundaries come from the constraint that the rear wheel slip angle is kept in the unsaturated region; considering that the lateral force of the rear wheel is saturated before the front wheel in the stability design of the vehicle, the maximum yaw rate r max : In the formula, μ is the adhesion coefficient of the driving road, ∝ is the tire side slip angle of the vehicle, g is the gravity acceleration, a and b are the lengths of the front and rear axles of the vehicle from the center of mass of the vehicle. -∝ max，r ≤∝ r ≤∝ max，r The trajectory tracking module specifically tracks the trajectory by the following AMPC method:

4. The system of claim 3, wherein: First, the vehicle dynamics model is simplified to the following form: ​ where Ccis the tire cornering stiffness; A cost function is designed, and a control amount deviation of the control step is obtained with the minimum cost function as an optimization target; the specific form of the cost function is as follows: J k = J zk + J uk + J duk + J εk where J zk , J uk , J duk are the cumulative cost function of tracking error, the cumulative cost function of control variable, the cumulative cost function of change of optimized control variable in the prediction time window, respectively, for ensuring the tracking of the vehicle to the target trajectory, J εk is the cost function of relaxation variable, which has the specific form as follows: Wherein, x(k+h|k) is the vehicle state at k+h| time, u(k+h) is the vehicle action performed at k+h| time, Q(k+h|k), R(k), E(k) are all weight matrices corresponding to k+h time, S z is a diagonal matrix of normalization coefficients, s u , s du are all normalization coefficients; The multi-objective optimization problem based on AMPC is represented as: where u ass,max is the maximum allowed correction, u max is the limit value of the actuator.

5. A multi-domain collaborative stability control method for intelligent electrically driven heavy-duty vehicles, characterized in that: The method is performed using the system of any of claims 1-4, and specifically includes the following steps: S1, using the vehicle sensing signals provided by the multiple types of sensors, the road adhesion coefficient estimation module estimates the road adhesion coefficient of the road on which the vehicle is currently located in real time, and the yaw rate module calculates the yaw rate; the road condition prediction module fuses the road adhesion coefficient estimation value with the visual sensor to obtain the reliable road adhesion coefficient of the front road section; S2, the safety boundary analysis module integrates the sensing signals and the data obtained in step S1, and represents the stability of the vehicle through a centric side slip angle-yaw rate phase plane; the vehicle yaw rate and centric side slip angle constraints under the maximum performance of the rear wheel tire are set as the stability constraints for the stability envelope of the vehicle attitude; S3, the steering control module and the speed control module establish the vehicle safety boundary based on the stability constraints obtained in S2, map the real-time state point of the vehicle, judge the stability state in which the vehicle is located and the stability risk at the future time, and output the steering and speed boundary values at which the vehicle can be stably driven; S4, the driving trajectory re-planning module re-plans the driving trajectory of the vehicle according to the current state point of the vehicle, the safety boundary, and the estimated road adhesion coefficient in front of the vehicle; S5, the trajectory tracking module tracks the driving trajectory of the vehicle in an AMPC manner, and controls the driving, braking, and steering systems of the vehicle.

Images