重庆理工大学学报2025，Vol.39Issue(19)：38-48,11.DOI:10.3969/j.issn.1674-8425(z).2025.10.005
融合最优避障路径的路径跟踪与横向稳定性协同博弈控制
Cooperative game control of path tracking and lateral stability with integrated optimal obstacle avoidance path
摘要
Abstract
The rapid development of autonomous driving technology markedly improves vehicles'intelligence and safety.Autonomous driving systems integrate multiple sensors,decision-making algorithms,and control mechanisms to achieve real-time perception,information fusion,and dynamic decision-making in complex environment.These systems generate control commands within milli-seconds,reduce dependence on human operation,and substantially lower risks caused by human errors.Thus,traffic efficiency and driving comfort are improved. Within the vehicle architecture,path planning and path-tracking control serve as the core of autonomous navigation.Path planning generates feasible,smooth,and safe trajectories while path-tracking control ensures precise and stable trajectory.The coordinated optimization of these processes directly determines vehicle performance and safety during real-time operation. Dynamic driving scenarios with moving obstacles impose demanding challenges on path planning and tracking control.Path planning quickly adjusts trajectories within limited time to avoid collisions while considering road boundaries,surrounding traffic participants,and vehicle dynamics and kinematic constraints.Algorithms that perform well in static environment often fail to maintain both dynamic feasibility and real-time efficiency in complex scenarios,which undermines performances.Conventional methods usually generate geometrically feasible trajectories but ignore vehicle dynamics,producing tracking errors and instability during high-speed lane changes or emergency maneuvers.Real-time obstacle avoidance further increases computation and lowers efficiency.Traditional path-tracking strategies emphasize positional accuracy but neglect yaw stability,which is essential for controllability during sharp turns or sudden avoidance maneuvers.Over-emphasis on accuracy generates excessive yaw rates and instability,whereas conservative strategies compromise tracking performance and raise collision risks.Effective path planning and control strategies therefore balance real-time adaptability,vehicle dynamics,tracking precision,and yaw stability to achieve safe and reliable autonomous driving in complex environment. A cooperative game-based control strategy integrates optimal obstacle-avoidance paths to address these challenges and enhance vehicle safety and control performance in high-speed dynamic scenarios.At the path planning level,a potential-field-driven model predictive control(PF-MPC)local planner continuously generates smooth and optimized avoidance trajectories.This planner incorporates dynamic obstacle information,road boundaries,and global reference paths into a unified cost function,effectively modeling environmental influences through obstacle and boundary potential fields.Coordinate transformation and potential field convexification convert nonlinear constraints into a least-squares optimization problem,which improves computational efficiency and supports real-time performance.Polynomial fitting further smooths and optimizes trajectories.The tracking module executes them directly.At the control level,a dynamic Nash game-based coordination method resolves conflicts between tracking accuracy and yaw stability.The path-tracking controller and yaw-stabilization controller act as players in a dynamic game,each pursuing its own objective—one emphasizes trajectory-following accuracy while the other prioritizes lateral stability.A state-prediction and strategy-coupling framework enables both controllers to optimize cooperatively during the game process.Within a distributed model predictive control(DMPC)framework,solving the Nash equilibrium generates the globally optimal front-wheel steering sequence.Simulation evidence confirms the analytical solution delivers superior real-time performance compared with quadratic programming,while the Nash game-based strategy achieves better coordination between path-tracking accuracy and lateral stability than conventional predictive control approaches.关键词
路径规划/模型预测控制/纳什博弈/协同控制Key words
path planning/model predictive control/Nash game/cooperative control分类
交通运输引用本文复制引用
严运兵,李胜,郭慧婕,刘申奥..融合最优避障路径的路径跟踪与横向稳定性协同博弈控制[J].重庆理工大学学报,2025,39(19):38-48,11.基金项目
国家自然科学基金项目(51975428) （51975428）
