A composite delay estimation and compensation method for drive-by-wire chassis trajectory tracking control

Abstract

This invention discloses a steerable chassis trajectory tracking control method based on composite time delay estimation and compensation, comprising the following steps: establishing a discretized vehicle dynamics model considering steering execution lag, clarifying the difference between steering lag and stochastic network delay, and providing a basis for compensation design; designing an adaptive prediction time-domain function, dynamically calculating the optimal prediction time domain based on the current vehicle speed, road curvature, and network delay state; designing a control quantity considering steering lag and constructing a time delay uncertainty model, using the polyhedral method to handle the parameter uncertainty caused by time delay; performing rolling optimization solution, dynamically adjusting the optimization weights based on the uncertainty model, the adaptive time domain, and an online weight adaptive mechanism of reinforcement learning to obtain the optimal control increment. This invention solves the problems in existing technologies that simplify complex internal time delays into a single, fixed model, failing to distinguish and process their physical causes, resulting in limited compensation accuracy, poor system dynamic adaptability, and insufficient robustness.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, specifically to a drive-by-wire chassis trajectory tracking control method based on composite time delay estimation and compensation. Background Technology

[0002] New energy vehicles have become an important part of technological innovation and industrial upgrading in the automotive industry, fundamentally reducing energy consumption and environmental pollution. Chassis-based drive-by-wire technology is a technological development trend that meets the demands of intelligent electric vehicles for higher execution precision and faster response speeds, and is a fundamental guarantee for realizing intelligent vehicle applications and precise control.

[0003] With the rapid development of intelligent vehicles and intelligent driving technology, steer-by-wire systems, as a key vehicle chassis control technology, have gradually become an important direction for industry development. Steer-by-wire systems replace traditional mechanical steering connections with electronic signals to achieve precise control of steering actuators (such as motors, servos, etc.), which not only improves the controllability and flexibility of the whole vehicle, but also provides higher freedom and integration for advanced driver assistance systems and autonomous driving systems. However, there is a complex compound internal time delay between the upper controller issuing the command and the actual vehicle response, mainly including: (1) Actuator lag: Due to motor characteristics, gear backlash and friction damping, there is a physical lag in the steering actuator when responding to control commands. (2) Network communication delay: Vehicle sensors, controllers and actuators communicate through the CAN bus, and bandwidth limitations and data congestion will lead to random network delays.

[0004] In existing research on tracking control of steer-by-wire systems, for example, Chinese invention patent application number CN202110029797.8, entitled "A vehicle merging control system and method considering steer-by-wire hysteresis," analyzes the impact of steer-by-wire hysteresis on merging control, comprehensively considers the safety and comfort of the vehicle merging process, establishes a corresponding cost function, and solves for the optimal action at each moment during the vehicle merging process, thereby improving the safety and comfort of vehicle merging; Chinese invention patent application number CN202411518989.5, entitled "A steer-by-wire stability control method based on event triggering-H∞ robustness," solves the controller gain matrix based on an augmented system model, combines the event triggering signal and the reference model output, and generates an H∞ robust control input signal to solve the disturbance problems such as system parameter uncertainty and communication delay in steer-by-wire system control. Chinese invention patent application number CN202510971392.4, entitled "A steerable chassis vehicle test device and method", identifies the delay characteristics of steering actuators by using a mixture of step and sinusoidal excitation signals, and deploys a delay compensation model in a real vehicle environment. The invention proposes an integrated steerable chassis vehicle test device and method that integrates "delay identification-model injection-real vehicle verification", providing an engineering solution for the dynamic and rapid calibration and algorithm migration of steerable chassis actuators for autonomous driving.

[0005] Existing technologies generally treat latency as a centralized, fixed, or only random parameter, without fully considering the multi-source and heterogeneous nature of its internal latency. The compensation strategies adopted are not robust enough, and the performance degrades under complex dynamic conditions.

[0006] Therefore, there is an urgent need for a control method that can classify and identify the internal time delay of the drive-by-wire chassis, model it differently, and compensate it under a unified optimization framework, so as to fundamentally solve the negative impact of composite time delay on trajectory tracking accuracy and robustness. Summary of the Invention

[0007] The present invention proposes a method for trajectory tracking control of a drive-by-wire chassis with composite time delay estimation and compensation, which can at least solve one of the technical problems in the background art.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A method for trajectory tracking control of a drive-by-wire chassis with composite time delay estimation and compensation includes the following steps: S1. Establish a discretized vehicle dynamics model that considers steering execution lag, clarify the difference between steering lag and stochastic network delay, and provide a basis for compensation design. S2. Design an adaptive prediction time-domain function to dynamically calculate the optimal prediction time domain based on the current vehicle speed, road curvature, and network latency. S3. Design a control variable that takes into account steering lag and construct a time delay uncertainty model. Use the polyhedron method to handle the parameter uncertainty caused by time delay. S4. Perform rolling optimization to solve the problem. Based on the uncertainty model constructed in step S3 and the adaptive time domain designed in step S2, as well as the online weight adaptive mechanism of reinforcement learning, dynamically adjust the optimization weights to obtain the optimal control increment.

[0009] As a preferred embodiment of the composite time delay estimation and compensation method for track tracking of a drive-by-wire chassis described in this invention, step S1 specifically includes the following steps: S11. Establish a two-degree-of-freedom dynamic model of the vehicle based on the actual rotation angle. Based on the vehicle single-track model, the deviation between the vehicle's lateral motion state and the ideal reference trajectory is selected as the state variable in the inertial coordinate system. OXY Below, the error vector between the vehicle state and the reference trajectory is defined. The error state equation in continuous time is: (1) Among them, the control input is This represents the deviation in the angle control quantity; x and y These refer to the longitudinal and lateral displacements of the vehicle, respectively. and These are the vehicle's longitudinal and lateral speeds, respectively. and These represent the vehicle yaw angle and yaw rate, respectively, and the system matrix. A t and control matrix B t The specific format is as follows:

[0010]

[0011]

[0012] in, C f 、C r These are the stiffnesses of the front and rear wheels, respectively. m For the sprung mass of the vehicle, δ f For the front wheel steering angle, l f ，l r These are the distances from the center of mass to the front and rear axles, respectively. Iz For vehicles to bypass z Moment of inertia of the shaft; This is the partial derivative of the lateral motion state with respect to the longitudinal velocity; The partial derivative of the longitudinal motion state with respect to the longitudinal velocity; For the partial derivative of the yaw motion state with respect to the longitudinal velocity; S12. To introduce the impact of steering execution lag of the SbW system on trajectory tracking, a rack and pinion mechanism of the SbW system is established, considering the pure lag and first-order inertial lag in its displacement response: (2) in, For pure time delay, It is the first-order inertial time constant; This leads to the hysteresis dynamic equation for the actual wheel rotation angle: (3) in, This is the actual front wheel steering angle. To determine the desired front wheel steering angle; S13. Discretize the above continuous model and continuous lag model, and set the sampling period. T s ,get: (4) in, (5) in, T s Sampling time; This is the symbol for the inverse Laplace transform; s For the Laplace operator; To optimize control performance, a control increment is introduced. As a new control variable, construct an augmented state vector. The discrete augmented system equations are obtained as follows: (6) in, (7) The discretized hysteresis model is as follows: (8) in, To compensate for the lag in the turnaround time.

[0013] As a preferred embodiment of the composite time delay estimation and compensation method for track tracking of a drive-by-wire chassis described in this invention, step S2 specifically includes the following steps: S21. At each sampling time, the prediction time-domain and control time-domain parameters change the dimension of the prediction equation coefficient matrix, thus affecting the solution direction of the rolling optimization. The dynamic adjustment function of the prediction time-domain is defined as follows: (9) in, N p ( k )for k The predicted time domain length at any given moment; V x ( k )for k The longitudinal velocity of the vehicle at any given moment; C ( k )for k Road curvature at any given moment for k Network latency estimate at time; k 1, k 2, k 3, b The adaptive adjustment coefficient is _____; round(·) is the rounding function; S22, Define the control time domain N c The relationship with the prediction time domain is as follows: (10) in, for k The control time domain length at any given moment; n To control the ratio of the time domain to the prediction time domain; S23. Adaptive parameters are determined through offline optimization, and the optimized model is: (11) in, This is the overall loss function; Optimize computation time for single-step scrolling; The average trajectory tracking error over the entire process; , These are the weighting coefficients. The physical meaning of the constraint is that the CPU computation time required for single-step rolling optimization should not exceed the sampling time.

[0014] As a preferred embodiment of the composite time delay estimation and compensation method for track tracking of a drive-by-wire chassis described in this invention, step S3 specifically includes the following steps: S31. Considering the steering lag caused by mechanical backlash and friction damping in the steering actuator, establish a control quantity model containing the lag dynamics; define the actual control vector. Desired control vector According to equation (8) in step S13, the control quantity considering steering lag is expressed as: (12) Discretizing equation (12) yields: (13) S32. Considering random delay in CAN network The augmented state vector contains historical control variables: (14) in, for k Network latency at any given moment; γ This represents the maximum number of delay steps. for ki The amount of control at any given moment; The model for a stochastic time-delay uncertain system is established as follows: (15) in,

[0015]

[0016] The uncertainty coefficient is defined as: (16) (17) in, It is the upper limit of integration, and its physical meaning is the limit of the controller node's integration. ki Signals sent at all times Due to network latency extending to the first k The time interval corresponding to the moment; Is the controller node in ki The total duration of messages sent at any given time; S33. The polyhedral representation of the time delay uncertainty coefficient is established as follows: (18) in, h The number of vertices of the polyhedron; These are time-varying weighting coefficients; This is the uncertainty matrix corresponding to the vertices of the polyhedron; Define the vertices of the polyhedron as: (19) in, , To control the proportion of effective time that messages spend in computation; The uncertainty coefficient matrix of time delay in a random network is represented by a polyhedron as follows: (20).

[0017] As a preferred embodiment of the composite time delay estimation and compensation method for track tracking of a drive-by-wire chassis described in this invention, step S4 specifically includes the following steps: S41. The predicted output equation is constructed as follows: (twenty one) in, To predict the output sequence; To control the incremental sequence; and The coefficient matrices for the state variables and control variables are shown below: S42. Integrate the vehicle's real-time motion state, tracking error, environmental information, and system latency characteristics to construct a state vector. s ( k As input: (twenty two) in, V x For the longitudinal speed of the vehicle, e y and e φ These are the lateral position error and the yaw angle error, respectively. and The rate of change of error, This is an estimate of network latency. This represents the standard deviation of the recent tracking error. To optimize the solution time for the previous control cycle; Secondly, define the action vector. a ( k Let be a two-dimensional continuous variable, representing the scaling factor of the weight matrix: (twenty three) in, State deviation weight matrix Q The scaling factor; Weight matrix Q The upper and lower limits of the scaling factor; State deviation weight matrix R The scaling factor; Weight matrix R The upper and lower limits of the scaling factor; Controller in kThe time-varying weight matrix actually used at any given time is determined by the following formula: (twenty four) in, and This is the baseline weight matrix obtained through system calibration; Furthermore, design the reward function. r ( k To comprehensively evaluate single-step control performance: (25) in, to These are the weighting coefficients. (·) is an indicator function, used to optimize computation time. t c Exceeding the sampling period T s The value is 1 if the condition is met, and 0 otherwise. The Actor-Critic framework is used, and online training is performed through a deep deterministic policy gradient algorithm, based on the current state. s ( k Output Action a ( k Rewards are given after interacting with the environment. r ( k And update the policy network parameters to achieve continuous optimization of the weight adaptive policy; S43. Based on the adaptive weight matrix obtained in step S42 At each sampling time k To solve for the constrained finite-time optimal control, define the rolling optimization objective function: (26) (27) in, for k+i Reference output at any given moment; The weights are relaxation factor weights; These are slack variables; To control the upper and lower limits of incremental constraints; To control the upper and lower limits of the quantity; To output the upper and lower limits of hard constraints; To output the upper and lower limits of soft constraints; The optimization problem is then transformed into a standard quadratic programming form: (28) in, To optimize the variable vector; It is a Hessian matrix; The gradient vector; This is the prediction error vector; For the constraint matrix, To constrain the upper and lower limits, we obtain from equation (27); blkdiag(⋅) is the block diagonal matrix constructor; The optimal control increment sequence is obtained by solving. Extract the control increment at the current moment: (28) in, To control the input dimensions; Update the control variable: (29) It is sent as the desired steering angle command to the steer-by-wire system to complete the steer-by-wire chassis trajectory tracking control to compensate for the steering execution lag and random network delay within the chassis.

[0018] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0019] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0020] The beneficial effects of this invention are: This invention performs mechanism analysis and differentiated modeling of steering mechanical lag and stochastic network delay within a steer-by-wire chassis within a unified control framework. It also innovatively integrates a lag compensation method based on state augmentation and an uncertainty handling strategy based on the polyhedron method, thereby achieving estimation and compensation for two different types of delays.

[0021] This invention designs an adaptive strategy that can dynamically adjust the prediction time domain according to real-time operating conditions, and supplements it with offline parameter optimization. This enables the system to maintain high-precision trajectory tracking and strong robust stability under various complex conditions such as low-adhesion road surfaces, variable vehicle speeds, and variable loads, while strictly meeting real-time constraints.

[0022] This invention fills the gap in existing technologies for handling heterogeneous time delays within steerable drive chassis, proposing a complete control architecture from modeling and compensation to optimization. This invention provides a reliable technical solution for accurate trajectory tracking of steerable drive chassis under complex working conditions, possessing significant engineering application and promotional value. Attached Figure Description

[0023] Picture 1This is a schematic diagram of the wire-controlled chassis trajectory tracking control method based on composite time delay estimation and compensation according to the present invention.

[0024] Picture 2 This is a schematic diagram of the polyhedral time delay uncertainty of the wire-controlled chassis trajectory tracking control method of the present invention, which combines time delay estimation and compensation.

[0025] Picture 3 This is a diagram of the reinforcement learning Actor-Critic structure for the composite time delay estimation and compensation method for trajectory tracking control of a drive-by-wire chassis according to the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0027] like Picture 1-Picture 3 As shown in the figure, a method for trajectory tracking control of a drive-by-wire chassis with composite time delay estimation and compensation in this embodiment includes the following steps: S1. Establish a discretized vehicle dynamics model that considers steering execution lag, clarify the difference between steering lag and stochastic network delay, and provide a basis for compensation design. S2. Design an adaptive prediction time-domain function to dynamically calculate the optimal prediction time domain based on the current vehicle speed, road curvature, and network latency. S3. Design a control variable that takes into account steering lag and construct a time delay uncertainty model. Use the polyhedron method to handle the parameter uncertainty caused by time delay. S4. Perform rolling optimization to solve the problem. Based on the uncertainty model constructed in step S3 and the adaptive time domain designed in step S2, as well as the online weight adaptive mechanism of reinforcement learning, dynamically adjust the optimization weights to obtain the optimal control increment.

[0028] Specifically, step S1 includes the following steps: S11. Establish a two-degree-of-freedom dynamic model of the vehicle based on the actual rotation angle. Based on the vehicle single-track model, the deviation between the vehicle's lateral motion state and the ideal reference trajectory is selected as the state variable in the inertial coordinate system. OXY Below, the error vector between the vehicle state and the reference trajectory is defined. The error state equation in continuous time is: (1) Among them, the control input is This represents the deviation in the angle control quantity; x and y These refer to the longitudinal and lateral displacements of the vehicle, respectively. and These are the vehicle's longitudinal and lateral speeds, respectively. and These represent the vehicle yaw angle and yaw rate, respectively, and the system matrix. A t and control matrix B t The specific format is as follows:

[0029]

[0030]

[0031] in, C f 、C r These are the stiffnesses of the front and rear wheels, respectively. m For the sprung mass of the vehicle, δ f For the front wheel steering angle, l f ，l r These are the distances from the center of mass to the front and rear axles, respectively. I z For vehicles to bypass z Moment of inertia of the shaft; This is the partial derivative of the lateral motion state with respect to the longitudinal velocity; The partial derivative of the longitudinal motion state with respect to the longitudinal velocity; For the partial derivative of the yaw motion state with respect to the longitudinal velocity; S12. To introduce the impact of steering execution lag of the SbW system on trajectory tracking, a rack and pinion mechanism of the SbW system is established, considering the pure lag and first-order inertial lag in its displacement response: (2) in, For pure time delay, It is the first-order inertial time constant; This leads to the hysteresis dynamic equation for the actual wheel rotation angle: (3) in, This is the actual front wheel steering angle. To determine the desired front wheel steering angle; S13. Discretize the above continuous model and continuous lag model, and set the sampling period. Ts ,get: (4) in, (5) in, T s Sampling time; This is the symbol for the inverse Laplace transform; s For the Laplace operator; To optimize control performance, a control increment is introduced. As a new control variable, construct an augmented state vector. The discrete augmented system equations are obtained as follows: (6) in, (7) The discretized hysteresis model is as follows: (8) in, To compensate for the lag in the turnaround time.

[0032] Furthermore, step S2 specifically includes the following steps: S21. At each sampling time, the prediction time-domain and control time-domain parameters change the dimension of the prediction equation coefficient matrix, thus affecting the solution direction of the rolling optimization. The dynamic adjustment function of the prediction time-domain is defined as follows: (9) in, N p ( k )for k The predicted time domain length at any given moment; V x ( k )for k The longitudinal velocity of the vehicle at any given moment; C ( k )for k Road curvature at any given moment for k Network latency estimate at time; k 1, k 2, k 3, b The adaptive adjustment coefficient is _____; round(·) is the rounding function; S22, Define the control time domain N c The relationship with the prediction time domain is as follows: (10) in, for k The control time domain length at any given moment; n To control the ratio of the time domain to the prediction time domain; S23. Adaptive parameters are determined through offline optimization, and the optimized model is: (11) in, This is the overall loss function; Optimize computation time for single-step scrolling; The average trajectory tracking error over the entire process; , These are the weighting coefficients. The physical meaning of the constraint is that the CPU computation time required for single-step rolling optimization should not exceed the sampling time.

[0033] Step S3 specifically includes the following steps: S31. Considering the steering lag caused by mechanical backlash and friction damping in the steering actuator, establish a control quantity model containing the lag dynamics; define the actual control vector. Desired control vector According to equation (8) in step S13, the control quantity considering steering lag is expressed as: (12) Discretizing equation (12) yields: (13) S32. Considering random delay in CAN network The augmented state vector contains historical control variables: (14) in, for k Network latency at any given moment; γ This represents the maximum number of delay steps. for ki The amount of control at any given moment; The model for a stochastic time-delay uncertain system is established as follows: (15) in,

[0034]

[0035] The uncertainty coefficient is defined as: (16) (17) in, It is the upper limit of integration, and its physical meaning is the limit of the controller node's integration. ki Signals sent at all times Due to network latency extending to the first k The time interval corresponding to the moment; Is the controller node in ki The total duration of messages sent at any given time; S33. The polyhedral representation of the time delay uncertainty coefficient is established as follows: (18) in, h The number of vertices of the polyhedron; These are time-varying weighting coefficients; This is the uncertainty matrix corresponding to the vertices of the polyhedron; Define the vertices of the polyhedron as: (19) in, , To control the proportion of effective time that messages spend in computation; The uncertainty coefficient matrix of time delay in a random network is represented by a polyhedron as follows: (20).

[0036] Furthermore, step S4 specifically includes the following steps: S41. The predicted output equation is constructed as follows: (twenty one) in, To predict the output sequence; To control the incremental sequence; and The coefficient matrices for the state variables and control variables are shown below: S42. Integrate the vehicle's real-time motion state, tracking error, environmental information, and system latency characteristics to construct a state vector. s ( k As input: (twenty two) in, V x For the longitudinal speed of the vehicle, e y and e φ These are the lateral position error and the yaw angle error, respectively. and The rate of change of error, This is an estimate of network latency. This represents the standard deviation of the recent tracking error. To optimize the solution time for the previous control cycle; Secondly, define the action vector. a ( k Let be a two-dimensional continuous variable, representing the scaling factor of the weight matrix: (twenty three) in, State deviation weight matrix Q The scaling factor; Weight matrix Q The upper and lower limits of the scaling factor; State deviation weight matrix R The scaling factor; Weight matrix R The upper and lower limits of the scaling factor; Controller in k The time-varying weight matrix actually used at any given time is determined by the following formula: (twenty four) in, and This is the baseline weight matrix obtained through system calibration; Furthermore, design the reward function. r ( k To comprehensively evaluate single-step control performance: (25) in, to These are the weighting coefficients. (·) is an indicator function, used to optimize computation time. t c Exceeding the sampling period T s The value is 1 if the condition is met, and 0 otherwise. The Actor-Critic framework is used, and online training is performed through a deep deterministic policy gradient algorithm, based on the current state. s ( k Output Action a ( k Rewards are given after interacting with the environment. r ( k And update the policy network parameters to achieve continuous optimization of the weight adaptive policy; S43. Based on the adaptive weight matrix obtained in step S42 At each sampling timek To solve for the constrained finite-time optimal control, define the rolling optimization objective function: (26) (27) in, for k+i Reference output at any given moment; The weights are relaxation factor weights; These are slack variables; To control the upper and lower limits of incremental constraints; To control the upper and lower limits of the quantity; To output the upper and lower limits of hard constraints; To output the upper and lower limits of soft constraints; The optimization problem is then transformed into a standard quadratic programming form: (28) in, To optimize the variable vector; It is a Hessian matrix; The gradient vector; This is the prediction error vector; For the constraint matrix, To constrain the upper and lower limits, we obtain from equation (27); blkdiag(⋅) is the block diagonal matrix constructor; The optimal control increment sequence is obtained by solving. Extract the control increment at the current moment: (28) in, To control the input dimensions; Update the control variable: (29) It is sent as the desired steering angle command to the steer-by-wire system to complete the steer-by-wire chassis trajectory tracking control to compensate for the steering execution lag and random network delay within the chassis.

[0037] This invention performs mechanism analysis and differentiated modeling of steering mechanical lag and stochastic network delay within a steer-by-wire chassis within a unified control framework. It also innovatively integrates a lag compensation method based on state augmentation and an uncertainty handling strategy based on the polyhedron method, thereby achieving estimation and compensation for two different types of delays.

[0038] This invention designs an adaptive strategy that can dynamically adjust the prediction time domain according to real-time operating conditions, and supplements it with offline parameter optimization. This enables the system to maintain high-precision trajectory tracking and strong robust stability under various complex conditions such as low-adhesion road surfaces, variable vehicle speeds, and variable loads, while strictly meeting real-time constraints.

[0039] This invention fills the gap in existing technologies for handling heterogeneous time delays within steerable drive chassis, proposing a complete control architecture from modeling and compensation to optimization. This invention provides a reliable technical solution for accurate trajectory tracking of steerable drive chassis under complex working conditions, possessing significant engineering application and promotional value.

[0040] In another aspect, the present invention also discloses a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the method described above.

[0041] In another aspect, the present invention also discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of the method described above.

[0042] In another embodiment provided in this application, a computer program product containing instructions is also provided, which, when run on a computer, causes the computer to execute any of the composite time delay estimation and compensation methods for drive-by-wire chassis trajectory tracking control described in the above embodiments.

[0043] It is understood that the systems, devices, and storage media provided in the embodiments of the present invention correspond to the methods provided in the embodiments of the present invention, and the explanations, examples, and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.

[0044] In the above embodiments, implementation can be achieved entirely or partially through software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented entirely or partially in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state disk (SSD)).

[0045] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0046] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0047] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for trajectory tracking control of a drive-by-wire chassis based on composite time delay estimation and compensation, characterized in that, Includes the following steps: S1. Establish a discretized vehicle dynamics model that considers steering execution lag, clarify the difference between steering lag and stochastic network delay, and provide a basis for compensation design. S2. Design an adaptive prediction time-domain function to dynamically calculate the optimal prediction time domain based on the current vehicle speed, road curvature, and network latency. S3. Design a control variable that takes into account steering lag and construct a time delay uncertainty model. Use the polyhedron method to handle the parameter uncertainty caused by time delay. S4. Perform rolling optimization to solve the problem. Based on the uncertainty model constructed in step S3 and the adaptive time domain designed in step S2, as well as the online weight adaptive mechanism of reinforcement learning, dynamically adjust the optimization weights to obtain the optimal control increment.

2. The method for trajectory tracking control of a drive-by-wire chassis based on composite time delay estimation and compensation according to claim 1, characterized in that: Step S1 specifically includes the following steps: S11. Establish a two-degree-of-freedom dynamic model of the vehicle based on the actual rotation angle. Based on the vehicle single-track model, the deviation between the vehicle's lateral motion state and the ideal reference trajectory is selected as the state variable in the inertial coordinate system. OXY Below, the error vector between the vehicle state and the reference trajectory is defined. The error state equation in continuous time is: (1) Among them, the control input is This represents the deviation in the angle control quantity; x and y These refer to the longitudinal and lateral displacements of the vehicle, respectively. and These are the vehicle's longitudinal and lateral speeds, respectively. and These represent the vehicle yaw angle and yaw rate, respectively, and the system matrix. A t and control matrix B t The specific format is as follows: ； ； ； ； ； in, C f 、C r These are the stiffnesses of the front and rear wheels, respectively. m For the sprung mass of the vehicle, δ f For the front wheel steering angle, l f ，l r These are the distances from the center of mass to the front and rear axles, respectively. I z For vehicles to bypass z Moment of inertia of the shaft; This is the partial derivative of the lateral motion state with respect to the longitudinal velocity; The partial derivative of the longitudinal motion state with respect to the longitudinal velocity; For the partial derivative of the yaw motion state with respect to the longitudinal velocity; S12. To introduce the impact of steering execution lag of the SbW system on trajectory tracking, a rack and pinion mechanism of the SbW system is established, considering the pure lag and first-order inertial lag in its displacement response: （2） in, For pure time delay, It is the first-order inertial time constant; This leads to the hysteresis dynamic equation for the actual wheel rotation angle: （3） in, This is the actual front wheel steering angle. To determine the desired front wheel steering angle; S13. Discretize the above continuous model and continuous lag model, and set the sampling period. T s ,get: （4） in, （5） in, T s Sampling time; This is the symbol for the inverse Laplace transform; s For the Laplace operator; To optimize control performance, a control increment is introduced. As a new control variable, construct an augmented state vector. The discrete augmented system equations are obtained as follows: （6） in, （7） The discretized hysteresis model is as follows: （8） in, To compensate for the lag in the turnaround time.

3. The method for trajectory tracking control of a drive-by-wire chassis based on composite time delay estimation and compensation according to claim 2, characterized in that: Step S2 specifically includes the following steps: S21. At each sampling time, the prediction time-domain and control time-domain parameters change the dimension of the prediction equation coefficient matrix, thus affecting the solution direction of the rolling optimization. The dynamic adjustment function of the prediction time-domain is defined as follows: （9） in, N p ( k )for k The predicted time domain length at any given moment; V x ( k )for k The longitudinal velocity of the vehicle at any given moment; C ( k )for k Road curvature at any given moment for k Network latency estimate at time; k 1, k 2, k 3, b The adaptive adjustment coefficient is _____; round(·) is the rounding function; S22, Define the control time domain N c The relationship with the prediction time domain is as follows: （10） in, for k The control time domain length at any given moment; n To control the ratio of the time domain to the prediction time domain; S23. Adaptive parameters are determined through offline optimization, and the optimized model is: （11） in, This is the overall loss function; Optimize computation time for single-step scrolling; The average trajectory tracking error over the entire process; , These are the weighting coefficients. The physical meaning of the constraint is that the CPU computation time required for single-step rolling optimization should not exceed the sampling time.

4. The method for trajectory tracking control of a drive-by-wire chassis based on composite time delay estimation and compensation according to claim 3, characterized in that: Step S3 specifically includes the following steps: S31. Considering the steering lag caused by mechanical backlash and friction damping in the steering actuator, establish a control quantity model containing the lag dynamics; define the actual control vector. Desired control vector According to equation (8) in step S13, the control quantity considering steering lag is expressed as: （12） Discretizing equation (12) yields: （13） S32. Considering random delay in CAN network The augmented state vector contains historical control variables: （14） in, for k Network latency at any given moment; γ This represents the maximum number of delay steps. for ki The amount of control at any given moment; The model for a stochastic time-delay uncertain system is established as follows: （15） in, ； ； The uncertainty coefficient is defined as: （16） （17） in, It is the upper limit of integration, and its physical meaning is the limit of the controller node's integration. ki Signals sent at all times Due to network latency extending to the first k The time interval corresponding to the moment; Is the controller node in ki The total duration of messages sent at any given time; S33. The polyhedral representation of the time delay uncertainty coefficient is established as follows: （18） in, h The number of vertices of the polyhedron; These are time-varying weighting coefficients; This is the uncertainty matrix corresponding to the vertices of the polyhedron; Define the vertices of the polyhedron as: （19） in, , To control the proportion of effective time that messages spend in computation; The uncertainty coefficient matrix of time delay in a random network is represented by a polyhedron as follows: （20）。 5. The method for trajectory tracking control of a drive-by-wire chassis based on composite time delay estimation and compensation according to claim 4, characterized in that: Step S4 specifically includes the following steps: S41. The predicted output equation is constructed as follows: （21） in, To predict the output sequence; To control the incremental sequence; and The coefficient matrices for the state variables and control variables are shown below: ； S42. Integrate the vehicle's real-time motion state, tracking error, environmental information, and system latency characteristics to construct a state vector. s ( k As input: （22） in, V x For the longitudinal speed of the vehicle, e y and e φ These are the lateral position error and the yaw angle error, respectively. and The rate of change of error, This is an estimate of network latency. This represents the standard deviation of the recent tracking error. To optimize the solution time for the previous control cycle; Secondly, define the action vector. a ( k Let be a two-dimensional continuous variable, representing the scaling factor of the weight matrix: （23） in, State deviation weight matrix Q The scaling factor; Weight matrix Q The upper and lower limits of the scaling factor; State deviation weight matrix R The scaling factor; Weight matrix R The upper and lower limits of the scaling factor; Controller in k The time-varying weight matrix actually used at any given time is determined by the following formula: （24） in, and This is the baseline weight matrix obtained through system calibration; Furthermore, design the reward function. r ( k To comprehensively evaluate single-step control performance: （25） in, to These are the weighting coefficients. (·) is an indicator function, used to optimize computation time. t c Exceeding the sampling period T s The value is 1 if the condition is met, and 0 otherwise. The Actor-Critic framework is used, and online training is performed through a deep deterministic policy gradient algorithm, based on the current state. s ( k Output Action a ( k Rewards are given after interacting with the environment. r ( k And update the policy network parameters to achieve continuous optimization of the weight adaptive policy; S43. Based on the adaptive weight matrix obtained in step S42 At each sampling time k To solve for the constrained finite-time optimal control, define the rolling optimization objective function: （26） （27） in, for k+i Reference output at any given moment; The weights are relaxation factor weights; These are slack variables; To control the upper and lower limits of incremental constraints; To control the upper and lower limits of the quantity; To output the upper and lower limits of hard constraints; To output the upper and lower limits of soft constraints; The optimization problem is then transformed into a standard quadratic programming form: （28） in, To optimize the variable vector; It is a Hessian matrix; The gradient vector; This is the prediction error vector; For the constraint matrix, To constrain the upper and lower limits, we obtain from equation (27); blkdiag(⋅) is the block diagonal matrix constructor; The optimal control increment sequence is obtained by solving. Extract the control increment at the current moment: （28） in, To control the input dimensions; Update the control variable: （29） It is sent as the desired steering angle command to the steer-by-wire system to complete the steer-by-wire chassis trajectory tracking control to compensate for the steering execution lag and random network delay within the chassis.

6. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, the processor performs the steps of the method as described in any one of claims 1 to 5.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the computer program is executed by the processor, it causes the processor to perform the steps of the method as described in any one of claims 1 to 5.

Images