智能汽车自适应RBF神经网络循迹控制

Adaptive RBF Neural Network Path Tracking Control for Intelligent Vehicle

为解决智能汽车循迹控制中建模复杂及不精确问题,提出了一种基于整体逼近的自适应RBF神经网络控制方法。首先,基于智能汽车动力学方程的基本形式,对系统的不确定性进行分析。而后,利用神经网络的逼近特性,对分析结果中的不确定项进行整体逼近。进而,基于自适应RBF神经网络控制方法设计控制律,并通过李雅普诺夫稳定性分析方法设计自适应控制律。最后,进行Simulink/Carsim联合仿真验证,仿真结果表明,在通用双移线道路环境下,所提控制方法能够通过控制方向盘转角使得车辆沿期望轨迹行驶,轨迹跟踪误差较小且控制输出幅值可控,能够满足实际工程需求。

In order to solve the problem of modeling complexity and imprecision in path tracking control of intelligent vehicle,an adaptive RBF neural network control method based on global approximation is proposed.Firstly,based on the basic form of intelligent vehicle dynamics equation,the uncertainty of the system is analyzed.Then,using the approximation characteristics of neural network,the uncertainty in the analysis results is approached as a whole.Then,the control law is designed based on the adaptive RBF neural network control method,and the adaptive control law is designed by the Lyapunov stability analysis method.Finally,the Simulink/CarSim joint simulation is carried out.The simulation results show that the proposed control method can make the vehicle follow the desired trajectory by controlling the steering wheel angle,and the trajectory tracking error is small and the control output amplitude is controllable,which can meet the actual engineering requirements.