农用车辆单神经元自适应PID轨迹跟踪控制  

Path Tracking of Agricultural Vehicles Based on Single Neuron Adaptive PID Control

作　　者：严友[1] 李美[2] YAN You;LI Mei(College of Electromechanical Engineering,Quzhou College of Technology,Zhejiang Quzhou 324000,China;College of Electromechanical Engineering,Hainan University,Hainan Haikou 570228,China)

机构地区：[1]衢州职业技术学院机电工程学院,浙江衢州324000 [2]海南大学机电工程学院,海南海口570228

出　　处：《机械设计与制造》2020年第10期228-231,235,共5页Machinery Design & Manufacture

基　　金：海南省自然科学基金项目(518QN209)。

摘　　要：为了实现复杂环境下的农用车辆路径跟踪控制,提出了基于单神经元自适应PID的路径跟踪控制方法。首先,建立了农用车辆运动学模型和电动助力转向系统模型,为后续的控制算法的设计打下模型基础。其次,设计了单神经元自适应PID控制策略,为了消除频繁控制引起的振荡,根据实际转角控制精度的要求,设置合理死区作为转角跟踪误差,利用带监督的赫步学习规则对神经元进行训练,在Matlab/Simulink环境下构建控制模型。试验和仿真结果表明,基于SNAPID的农用车辆路径跟踪控制精度较高,在直线行驶工况下,最大偏差在3.2cm以内,平均偏差在0.92cm以内;在曲线行驶工况下,最大偏差在4.3cm以内,平均偏差在1.03cm以内。In order to implement the path tracking control of agricultural vehicles under complex environment,a single neuron adaptive PID based path tracking control method was proposed.Firstly,the kinematic model of agricultural vehicles and the electric power steering system model were established,which laid the foundation for the design of the subsequent control algorithm.Secondly,a single neuron adaptive PID control strategy was designed.In order to eliminate the oscillation caused by frequent control,a reasonable dead zone was set as the corner tracking error according to the requirements of the actual steering control accuracy,and the use of supervised Hepburn learning rules was applied to neurons.The control model is built in the Matlab/Simulink environment.The results of experiments and simulations show that the SNA-PID based path tracking control for agricultural vehicles has higher precision.Under straight-line driving conditions,the maximum deviation is within3.2 cm,and the average deviation is within 0.92 cm.Under curve driving conditions,the maximum deviation is within 4.3 cm,the average deviation is within 1.03 cm.

关 键 词：农用车辆 轨迹跟踪 运动学模型 单神经元自适应 PID控制 

分 类 号：TH16[机械工程—机械制造及自动化] U461.6[机械工程—车辆工程]