中国工程机械学报

随着汽车智能化的迅猛发展，车道保持控制成为提升汽车驾驶安全性和舒适性的关键技术。为提升车辆运动轨迹跟踪精度，提出一种鲸鱼算法优化小波神经网络比例-积分-微分（PID）控制方法，通过仿真验证该方法对车辆横向位移误差的跟踪精度。构建车辆轮胎模型简图，定义车辆运动学方程式。引用小波神经网络PID控制方法，以提高控制的精度和适应性。为增强车辆控制系统的抗干扰能力，采用鲸鱼算法优化小波神经网络PID控制参数，使其能更好地逼近系统的动态特性，从而实现对车辆运动轨迹的有效控制。在不同行驶条件下，利用Matlab软件模拟车道保持效果。结果显示：采用小波神经网络PID控制系统，车辆按照参考运动轨迹行驶时，产生的横向误差较大；采用鲸鱼算法优化小波神经网络PID控制系统，车辆按照参考运动轨迹行驶时，产生的横向误差较小。所设计的鲸鱼算法优化小波神经网络PID控制系统，能够更精准地使车辆保持在既定的车道内，减少车道偏离。

With the rapid development of intelligent automobiles, lane keeping control has become a key technology for improving driving safety and comfort. In order to improve the tracking accuracy of vehicle motion trajectory, a whale algorithm optimized wavelet neural network PID control method is proposed, and the tracking accuracy of vehicle lateral displacement error is verified through simulation. Create a schematic diagram of the vehicle tire model and define the vehicle's kinematic equations. The wavelet neural network PID control method is referenced to improve the accuracy and adaptability of control. In order to enhance the disturbance capability of the vehicle control system, the whale algorithm is used to optimize the neural network PID control parameters, so that it can better approximate the dynamic characteristics of the system, thereby achieving effective control of the vehicle's motion trajectory. Simulate lane keeping effects using Matlab software under different driving conditions. The results show that when using the wavelet neural network PID control system, the lateral error generated by the vehicle traveling according to the reference motion trajectory is relatively large; By using the whale algorithm to optimize the wavelet neural network PID control system, the lateral error generated when the vehicle travels according to the reference motion trajectory is relatively small. The designed whale algorithm optimizes the wavelet neural network PID control system, which can more accurately keep the vehicle within the designated lane and reduce deviation.