Vehicle trajectory tracking control method and system capable of resisting parameter fluctuation and data packet loss

Abstract

The invention discloses a vehicle trajectory tracking control method and system capable of resisting parameter fluctuation and data packet loss, which are used for solving the problem that the control performance is reduced under the conditions of kinetic parameter fluctuation and data packet loss in the prior art. According to the method, a kinetic parameter uncertainty model is established through a bounded time-varying matrix, and a multi-modal system model comprising four sub-systems is constructed based on packet loss characteristics of a control instruction and state information. Dynamic behaviors of the system in different network states are uniformly described by adopting a Markov chain, the asymptotic stability of the system under the conditions of parameter disturbance and data packet loss is ensured by combining a Lyapunov stability theory and H-infinity control constraint, and the trajectory tracking precision and the driving stability are remarkably improved. In order to facilitate engineering realization, an original non-convex control problem is converted into a convex optimization problem which can be solved on line through a cone compensation linearization method. According to the method, the trajectory tracking robustness of the vehicle under the complex working condition is effectively improved, and the method has important engineering application value.

Description

Technical Field

[0001] This invention relates to the fields of automotive network communication and automotive control, and more specifically, to a vehicle trajectory tracking control method and system based on Markov chains that is resistant to parameter fluctuations and data packet loss. Background Technology

[0002] With the rapid development of autonomous driving technology, vehicle trajectory tracking control, as a key technology for achieving high-precision path following and safe driving, has become a research hotspot. However, in actual in-vehicle network environments, data packet loss is inevitable in communication links, and vehicle dynamics models generally exhibit time-varying and uncertain parameters. Both of these factors can significantly impact the performance and stability of trajectory tracking control systems. If not handled properly, they may even lead to control performance degradation and potential safety risks, severely restricting the reliability and safety of autonomous driving systems in complex dynamic environments.

[0003] Currently, trajectory tracking control methods for parameter uncertainty mainly fall into two categories: one is based on uncertain parameter compensation, which estimates changes in vehicle dynamic parameters online or offline and compensates for the control input accordingly to reduce the impact of model mismatch on control performance; the other is robust control strategies, which introduce parameter uncertainty models during the controller design phase to ensure the system maintains stability and control performance even under the most unfavorable conditions. Meanwhile, regarding the data packet loss problem, existing research has also proposed two main solutions: one is based on data compensation, which attempts to recover system performance by estimating or reconstructing lost data frames, but often struggles to balance estimation accuracy and computational complexity, limiting its real-time performance and practical feasibility; the other is based on robust control, which models packet loss as an uncertainty source and compensates for it to enhance system stability and anti-interference capabilities. However, most of these methods neglect the dynamic characteristics of the vehicle itself, resulting in insufficient robustness under complex conditions and failing to meet the control reliability and safety requirements of high-level autonomous driving systems.

[0004] While some studies have proposed solutions to the impact of parameter fluctuations or data loss on trajectory tracking control, most methods remain at the stage of treating the two types of problems separately, lacking the ability to unify modeling and coordinate responses. Furthermore, existing research largely focuses on offline modeling and simulation verification, lacking in-depth exploration of the effectiveness and engineering application value of control strategies in real-time online environments. Summary of the Invention

[0005] This invention addresses the performance degradation of existing vehicle trajectory tracking control methods when simultaneously affected by data loss and parameter fluctuations. It proposes a trajectory tracking control method and system with resilience against both parameter fluctuations and data loss. This method can simultaneously constrain the interference of dynamic parameter uncertainties and communication defects on closed-loop control performance, demonstrating significant potential for real-time applications.

[0006] Specifically, based on the vehicle's two-degree-of-freedom dynamics model and trajectory tracking kinematics model, by extracting key terms affected by parameter fluctuations, the original system is divided into a nominal fixed parameter model and a bounded time-varying parameter model, thereby clarifying the mechanism by which the uncertainty of dynamic parameters affects the vehicle's motion behavior and laying the modeling foundation for controller design.

[0007] Furthermore, considering the impact of data packet loss on system performance, the network state is divided into four typical scenarios based on its location in the forward and feedback channels: no packet loss, packet loss only in the forward channel, packet loss only in the feedback channel, and packet loss in both the forward and feedback channels simultaneously. Four subsystem models considering the effects of control signal and sensor signal loss are constructed, achieving accurate modeling of the dynamic changes in the system caused by packet loss and improving the completeness and accuracy of the model description.

[0008] Furthermore, to characterize the random transitions of network states, the evolution patterns of various network states were statistically analyzed, and an autocorrelation factor was introduced to describe the correlation between states at different times. Based on this, a Markov chain model was constructed, and a state transition probability matrix was established, thereby achieving a unified modeling of the network state evolution process and avoiding the frequent mode switching problem existing in traditional multi-model methods.

[0009] Furthermore, in terms of controller design, the following is introduced: The control method aims to suppress the impact of uncertain disturbances on control performance. Based on the Lyapunov stability principle, asymptotic stability conditions for Markov jump systems are constructed to ensure the stability and control performance of the system in multi-source disturbance environments.

[0010] Furthermore, to address the computational challenges posed by solving non-convex matrix inequalities, a cone-complement linearization method is employed to transform the problem into an iterative solution of linear matrix inequalities. By combining offline calculation of controller gain with online table lookup, the real-time performance of the control strategy is effectively improved, enabling the controller to be deployed on an online operating platform and meeting the real-time and robustness requirements of the vehicle control system. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art are briefly introduced below; obviously, the drawings in the following description are only embodiments of the present invention, and those skilled in the art can obtain other drawings based on the provided drawings without creative effort.

[0012] Figure 1 The overall framework diagram of the vehicle trajectory tracking control method and system based on anti-parameter fluctuation and data packet loss provided in the embodiments of the present invention is shown. Figure 2 This is a schematic diagram of a vehicle dynamics model provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of network data packet loss provided in an embodiment of the present invention. Detailed Implementation

[0013] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.

[0014] This invention addresses the uncertainties encountered during vehicle trajectory tracking, including fluctuations in dynamic parameters and communication data packet loss, by proposing a robust trajectory tracking control method and system. By introducing a bounded time-varying matrix to model parameter fluctuations and constructing a multimodal subsystem model under different communication packet loss scenarios, a unified modeling description of both types of uncertainties is achieved. Furthermore, a Markov chain is used to establish a transition mechanism between multiple subsystems, forming a unified switching system control framework. Simultaneously, a control law satisfying the Lyapunov stability condition is designed, and the non-convex optimization problem is transformed into an iteratively solvable linear matrix inequality problem using the cone complement linearization method, achieving both real-time performance and practicality of the controller.

[0015] like Figure 1 As shown, the vehicle trajectory tracking control method and system for resisting parameter fluctuations and data packet loss provided in this embodiment of the invention mainly includes two parts: a vehicle system modeling module and a robust controller design.

[0016] The specific steps for constructing a vehicle system model that integrates two types of uncertainty are as follows: S1. Considering the lateral control and heading angle control requirements of the vehicle in the trajectory tracking task, a two-degree-of-freedom lateral dynamics model and a trajectory tracking kinematic model of the vehicle are constructed to describe the vehicle's state response and trajectory evolution behavior.

[0017] S2, the uncertain parameter in the system is determined to be the front wheel lateral stiffness ( ), rear wheel lateral stiffness ( ) and longitudinal speed ( Based on the influence of these uncertain parameters on system dynamics, the overall system model is split into a nominal model with fixed parameters and a perturbation model with bounded time-varying parameters, thus constructing a vehicle dynamics description with structured uncertainty.

[0018] S3 models data packet loss in the forward channel as the loss of control input commands, and data packet loss in the feedback channel as the unavailability of vehicle state information. Based on different combinations of packet loss in the forward and feedback channels, the network state is divided into four typical scenarios, and four corresponding subsystem models are constructed for the system behavior under each network state.

[0019] S4. To characterize the switching patterns of the system between multiple modes, a Markov transition mechanism is introduced. Based on the historical statistical characteristics of the communication network states, a Markov state transition probability matrix is ​​constructed to model the transition behavior between different network states, thereby integrating the above-mentioned multiple subsystems into a unified Markov transition system.

[0020] Specifically, the vehicle dynamics and trajectory tracking kinematics model in step S1 are as follows: Figure 2 As shown, it can be expressed as the following equation: (1) in , , , and These are the sideslip angle, yaw angle, yaw rate, lateral position, and heading angle, respectively. , , and These are the vehicle's curb weight, and the surrounding area. The moment of inertia of the shaft and the distance between the front and rear shafts and the center of mass; and The front wheel steering angle is the additional yaw moment.

[0021] Specifically, the model in step S2 can be established through the following process: First, based on the influence of the uncertain parameters on the model, the model in equation (1) is transformed into the following form: (2) in, , , , , , , , , , , , , , and These are the lateral position and heading difference, respectively. and These are reference values ​​for lateral position and heading angle, respectively.

[0022] Will It is written as the nominal part of a fixed value and the time-varying fluctuation part, as shown below: (3) in The nominal value corresponds to , and When taking the nominal value Optional, and The nominal value is selected as 60,000 N / rad. The nominal value is selected based on the speed range. . For time-varying coefficients whose absolute value is less than or equal to 1. Then it is The range of variation of can be determined by the following formula: (4) The value depends on , and The range of variation, specifically within each speed range. The nominal value fluctuates within 2.5 m / s, which is optional. and It fluctuates within a range of 30,000 N / rad above and below the nominal value.

[0023] Will and In Replace with ,available and The nominal values ​​are respectively and ,definition , , , , , .

[0024] The system model can then be expressed in the following form: (5) in, , , , , , , , , To control the cycle.

[0025] Specifically, according to Figure 3 The network status in step S3 is shown in Table 1: Table 1 Network Status Classification Furthermore, four subsystem models can be constructed as follows: (6) in, , For the vehicle status to be successfully transmitted to the controller, , , , , This is the state feedback matrix.

[0026] Specifically, the unified model based on Markov chains in step S4 is as follows: (7) in Its change pattern can be observed through the state transition matrix. describe, Represents probability. This indicates that the system at any given time has a state. Jump to status The probability that satisfies .

[0027] The steps for building a controller that is resistant to parameter fluctuations and data packet loss are as follows: S1, determine the objective function for trajectory tracking and Performance metric functions; S2, based on the objective function of step S1 and Performance index functions are used to determine the conditions that guarantee system stability. S3 uses the cone complement linearization method to solve the control problem offline and obtains a controller parameter lookup table (MAP) that can be called online.

[0028] Specifically, regarding the performance index function in step S1, considering the requirements of minimizing trajectory tracking error, suppressing vehicle sideslip, and reducing control input amplitude, its performance function can be expressed as follows: (8) in and This is the weight matrix.

[0029] In step S1 The performance metric function is expressed as follows: (9) in , Defined Sexual indicators.

[0030] Specifically, the system stability conditions in step S2 are as follows: (10) (11) in, , , , , , , , , , , , , , , , , , , , , , In the above formula , , and The parameters to be determined.

[0031] Specifically, the offline solution and online application in step S3 are implemented in the following ways: Since equation (11) in step S2 is an equality constraint, the constructed control problem exhibits non-convex characteristics. To achieve convexity transformation of the problem, this equality constraint needs to be converted into a linear matrix inequality of the following form: (12) (13) Furthermore, based on the cone complement linearization method, the state feedback matrix of the online application can be obtained offline. MAP.

[0032] Based on the vehicle speed obtained from the feedback It can be determined superscript The value is It can be determined according to the following formula. superscript The value is (14) in The online state feedback matrix is ​​obtained by dynamically estimating the real-time network state through statistical analysis, and thus can be derived. .

[0033] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. The same or similar parts between the various embodiments can be referred to each other.

[0034] Note that the above embodiments are merely preferred embodiments of the present invention. The above description of the disclosed embodiments enables those skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A vehicle trajectory tracking control method and system resistant to parameter fluctuations and data packet loss, characterized in that, Applied to the field of autonomous driving control in automobiles, the described methods include: A parameter uncertainty modeling method based on bounded time-varying matrices is adopted to model the parameter fluctuations in the vehicle dynamics model; Construct a multimodal system model for different data packet loss scenarios to accurately describe the impact of communication anomalies on the dynamics of the closed-loop system; A unified modeling architecture for Markov chains that integrates parameter uncertainty and data loss is used to characterize the dynamic behavior of the system under random mode switching. A robust trajectory tracking control method for data loss and parameter disturbances is proposed to improve the robustness and accuracy of the control system under uncertain environments. The non-convex optimization problem in controller design is solved based on the cone complement linearization method, and the control strategy is efficiently deployed by combining offline iteration and online table lookup.

2. The method according to claim 1, characterized in that, In the vehicle modeling process, the system model is represented as the sum of the nominal parameter model and the fluctuation parameter model, thereby introducing the impact of parameter uncertainty on the trajectory tracking control performance.

3. The method according to claim 1, characterized in that, By considering data packet loss behavior in the forward channel (controller to actuator) and the feedback channel (sensor to controller), the network communication state is divided into four typical modes, and corresponding system models are established for each mode to achieve accurate modeling of the impact of data packet loss.

4. The method according to claim 1, characterized in that... Markov chains are used to model the transitions between network states, and a state transition probability matrix is ​​constructed to characterize the random switching behavior between communication modes, forming a unified modeling architecture that facilitates controller design and analysis.

5. The method according to claim 1, characterized in that, Based on Lyapunov stability theory and The control method involves constructing linear matrix inequality constraints to ensure the asymptotic stability of the Markov jump system, thereby suppressing uncertain disturbances introduced by data packet loss and ensuring high-precision trajectory tracking of the vehicle in unstable network environments.

6. The method according to claim 1, characterized in that, The non-convex matrix inequality problem is transformed into an iteratively solvable semidefinite programming problem by using the cone complement linearization method. The solution process is divided into two stages: offline iteration and online table lookup, so as to realize the rapid invocation and efficient deployment of control strategies in actual operation.

Images