CN-121979062-A - Dynamic permission distribution system and method oriented to man-machine sharing control

CN121979062ACN 121979062 ACN121979062 ACN 121979062ACN-121979062-A

Abstract

The invention provides a dynamic permission distribution system and method for man-machine sharing control, comprising the steps of establishing a vehicle longitudinal-transverse coupling dynamics model, constructing robust safety constraint and performance constraint, wherein the robust safety constraint is generated through an environment robust control barrier function, the environment robust control barrier function compensates the influence of sensor measurement uncertainty through a residual error correction term, the performance constraint is generated through a control Lyapunov function, the optimization problem is constructed based on the robust safety constraint and the performance constraint, the change rate of control permission of an automatic system is taken as a core decision variable, the robust safety constraint is set as a hard constraint, the performance constraint is set as a soft constraint, the optimization target is simultaneously minimized in permission fluctuation degree and control error, the optimization problem is solved, the current control permission weight is obtained, and a driver control instruction and an automatic system control instruction are subjected to weighted fusion to generate a final vehicle control instruction.

Inventors

CHEN YUTAO
ZHANG HONGLIANG
CHENG JUN
ZHOU YU
ZHENG FENG

Assignees

福州大学

Dates

Publication Date: 20260505
Application Date: 20260203

Claims (10)

1. A dynamic authority allocation method facing man-machine sharing control is characterized by comprising the following steps: Acquiring vehicle state information, a driver control instruction and an automation system control instruction; establishing a vehicle longitudinal-transverse coupling dynamics model for representing interaction between transverse and longitudinal motions of a vehicle; based on the coupling dynamics model, constructing a robust safety constraint and a performance constraint, wherein the robust safety constraint is generated through an environment robust control barrier function, and the environment robust control barrier function compensates the influence of the measurement uncertainty of the sensor through a residual error correction term; based on the robust safety constraint and the performance constraint, constructing an optimization problem by taking the change rate of the control authority of the automatic system as a core decision variable, setting the robust safety constraint as a hard constraint and the performance constraint as a soft constraint, and simultaneously minimizing the authority fluctuation degree and the control error as an optimization target; and solving the optimization problem to obtain the current control authority weight, and carrying out weighted fusion on the driver control instruction and the automatic system control instruction to generate a final vehicle control instruction.
2. The method for distributing dynamic rights oriented to man-machine sharing control according to claim 1, wherein the coupling dynamics model is a nonlinear affine dynamics model constructed by adopting a simplified bicycle model, comprises state equations of longitudinal position, transverse position, orientation and speed of a vehicle, and considers coupling influence of a slip angle and rolling resistance on motion of the vehicle, wherein the slip angle is calculated based on the distance from a vehicle mass center to a front shaft, the distance from the vehicle mass center to a rear shaft and a steering angle, and the rolling resistance is represented by adopting a polynomial model comprising a zero-order term, a first-order speed term and a second-order speed term.
3. The method for distributing dynamic rights oriented to man-machine sharing control according to claim 1 is characterized in that the construction process of the environment robust control barrier function comprises the steps of conducting quantitative analysis on uncertainty of a sensor measurement environment state, determining an error boundary, calculating deviation amount of a safety function and gradient of the safety function based on the error boundary, and integrating the deviation amount into a nominal safety constraint condition as a residual error correction term to form robust safety constraint resistant to sensor measurement errors.
4. The method for distributing dynamic permission for man-machine sharing control according to claim 1, wherein the performance targets of the control Lyapunov function comprise a transverse performance target and a longitudinal performance target, the transverse performance target is a square deviation of a transverse position of a vehicle from a reference transverse position, a square deviation of a vehicle orientation from a reference orientation, and the longitudinal performance target is a square deviation of a vehicle speed from a reference speed.
5. The method for distributing dynamic rights oriented to man-machine sharing control according to claim 1, wherein the optimization problem is a convex quadratic programming problem, the change rate of the rights is the difference between the control right weight at the current moment and the control right weight at the previous moment, the optimization target realizes control of the right fluctuation degree by minimizing the change rate, and when the convex quadratic programming problem is constructed, a relaxation variable is introduced to adjust the soft constraint so as to ensure the solvability of the optimization problem in a complex scene.
6. The method for distributing dynamic rights oriented to man-machine sharing control according to claim 5, wherein independent convex quadratic programming problems are respectively established for transverse control and longitudinal control of a vehicle, transverse control right weights and longitudinal control right weights are respectively solved through two independent optimizers, and the optimizers respond to control instructions of transverse/longitudinal dimension time asynchronism and ensure that control instructions of at least one dimension are effectively executed at any moment.
7. The method for distributing dynamic rights oriented to man-machine sharing control according to claim 1, wherein the value range of the control right weight is 0-1, and the variation of the control right weight is non-negative and does not exceed a preset maximum variation.
8. The method for dynamically distributing rights for man-machine sharing control according to claim 2, wherein the method is characterized in that the vehicle is controlled to execute the transverse auxiliary lane change and the longitudinal self-adaptive cruise based on the control instruction after the rights distribution coefficient is fused.
9. The method for dynamically assigning rights to human-machine-oriented shared control of claim 5, wherein said relaxation variables are used to ensure strict compliance of said robust security constraint by adjusting the satisfaction of said performance constraint when said performance constraint conflicts with said robust security constraint.
10. A human-machine shared vehicle dynamic rights allocation system, comprising: The information acquisition module is used for acquiring vehicle state information, a driver control instruction and an automatic system control instruction; the modeling module is used for establishing a vehicle longitudinal-transverse coupling dynamics model and representing interaction between transverse and longitudinal movements of the vehicle; The constraint construction module is used for constructing robust safety constraint through an environment robust control barrier function and performance constraint through control of a Lyapunov function based on the coupling dynamics model, and the robust safety constraint compensates sensor measurement uncertainty through a residual error correction term; the optimization module is used for constructing a convex quadratic programming optimization problem by taking the change rate of the control authority of the automatic system as a core decision variable based on the robust safety constraint and the performance constraint, setting the robust safety constraint as a hard constraint and the performance constraint as a soft constraint, and optimizing the target to minimize the authority fluctuation degree and the control error; and the instruction fusion module is used for solving the convex quadratic programming optimization problem to obtain control authority weight, and carrying out weighted fusion on the driver control instruction and the automatic system control instruction to generate a final vehicle control instruction.

Description

Dynamic permission distribution system and method oriented to man-machine sharing control Technical Field The invention belongs to the technical field of intelligent vehicle control and man-machine hybrid enhancement intelligence, and particularly relates to a dynamic permission distribution system and method for man-machine sharing control. Background Along with the development of intelligent driving technology, vehicle automation obviously improves road safety and traffic efficiency by reducing manual operation errors and optimizing driving tracks. However, in dynamic, complex open road scenarios, existing systems still face fundamental challenges such as environmental perception uncertainty, real-time decision constraints, etc. In this context, human-machine shared control (HMSC) has become a key way to achieve fully automated driving, combining human decision-making capability with control accuracy of the machine. The sharing control can be classified into direct and indirect sharing control according to the way in which the final control amount is synthesized. In direct shared control, human and system inputs together directly affect the vehicle's actuators, enabling real-time collaboration. However, direct shared control presents a potential risk of the driver competing with the automation system for control, possibly resulting in mutual interference and control instability. Currently, a key challenge in shared control research is the dynamic allocation mechanism of human-machine control rights. The existing research method mainly has the following limitations that a fixed weight distribution strategy is simple to realize but lacks clear physical meaning and adaptability, a distribution method based on a predefined index can provide a certain interpretation, but the effect of the distribution method is seriously dependent on a measurement standard design which is difficult to optimize, a quantization method based on a driving risk field tries to establish a mapping relation between environmental risks and authority distribution, but has defects in multi-source information fusion, and a method for modeling a shared driving process as a game theory problem faces the challenges of computational complexity caused by non-convex nonlinear optimization. The distribution method based on control Lyapunov function-control barrier function-quadratic programming (CLF-CBF-QP) appears in recent years, and the distribution method shows good safety, interpretability and real-time performance in a typical channel switching scene, thereby providing a new solution idea for dynamic authority distribution. However, this approach has yet to be perfected in terms of longitudinal-lateral cooperative control and sensor uncertainty handling. However, the above prior art has two major limitations in that most studies focus only on simplified lateral rights allocation, using steering angles as human input, ignoring the effects of longitudinal speed variation and longitudinal-lateral coupling, and most studies assume ideal perception of the environment, reducing the practical significance of the developed approach. Disclosure of Invention Aiming at the defects and shortcomings in the prior art, the invention provides a system and a method for distributing dynamic rights of a human-computer sharing vehicle, and belongs to the technical field of intelligent vehicle control. The scheme aims to solve the technical problems that the existing sharing control method only focuses on transverse control and ignores influence of sensor uncertainty and unsmooth authority transition, and constructs a dynamic authority distribution frame which takes longitudinal and transverse coordination and robust security into consideration. The method is characterized by comprising the steps of firstly establishing a longitudinal-transverse coupled vehicle nonlinear dynamics model, uniformly representing interaction between transverse steering and longitudinal speed change, providing an accurate model foundation for collaborative authority allocation, secondly designing an environment robust control barrier function (ER-CBF) aiming at sensor measurement uncertainty, constructing robust safety constraint tolerant to perceived errors by quantifying sensor error boundaries and introducing residual error correction items, simultaneously guaranteeing system control performance and stability by combining with Control Lyapunov Function (CLF), converting authority allocation problems into convex Quadratic Programming (QP) optimization problems, setting the robust safety constraint as hard constraint and performance constraint as soft constraint by taking the change rate of automatic system control authorities as a core decision variable, introducing relaxation variable to ensure the resolvability of the optimization problem under a complex scene, realizing smooth dynamic transition of authorities, and further establishing independent QP optimization problems for