基于ANFIS的四轮转向汽车控制策略研究  

Research on Control Strategy of Four-Wheel Steering Vehicle Based on ANFIS

作　　者：张俊 ZHANG Jun(Jinneng Holding Coal Industry Group Caoduogou Coal Industry Co.,Ltd,Datong Shanxi 037102,China)

机构地区：[1]晋能控股煤业集团草垛沟煤业有限公司,山西大同037102

出　　处：《机械研究与应用》2022年第5期165-168,173,共5页Mechanical Research & Application

摘　　要：针对前轮转向(2WS)汽车低速转弯半径大,高速稳定性差的问题,以二自由度四轮转向(4WS)车辆为研究对象,建立了汽车动力学模型和运动方程,结合模糊控制与人工神经网络控制形成自适应神经模糊推理(ANFIS)。以前轮转角和车速作为输入变量,后轮转角为输出值,建立包含训练样本的仿真实验模型,基于混合法训练得到自适应神经模糊推理控制器,在MATLAB/Simulink中建模,然后在低速和高速条件下,将自适应神经模糊推理控制器与前轮转向、前轮比例控制4WS、横摆角速度反馈控制4WS这三种控制策略进行仿真对比分析。仿真结果表明:自适应神经模糊控制使车辆在低速和高速时质心侧偏角趋近于零,横摆角速度和侧向加速度更加接近前轮转向车辆,具备优异的转向效果,极大地提高了车辆的行驶稳定性。Aiming at the problems of large turning radius at low speed and poor stability at high speed of front-wheel steering(2WS)vehicle,taking the two-degree-of-freedom four-wheel steering(4WS)vehicle as the research object,the vehicle dynamics model and motion equation are established,and the fuzzy control and artificial neural network control are combined to form Adaptive Neuro-Fuzzy Inference(ANFIS).The front wheel turning angle and vehicle speed are used as the input variables,and the rear wheel turning angle as the output value;then the simulation experimental model containing training samples is established.Based on the hybrid method training,an adaptive neuro-fuzzy inference controller is obtained and then modeled in MATLAB/Simulink;then,the three control strategies of the front wheel steering,front wheel proportional control 4WS and yaw rate feedback control 4WS are simulated and analyzed at low speed and high speed respectively.The simulation results show that the adaptive neuro-fuzzy control makes the vehicle's centroid angle close to zero at low and high speeds,and the yaw angular velocity and lateral acceleration are closer to the front-wheel steering vehicle.Thus it has excellent steering effect,and greatly improves the driving stability of the vehicle.

关 键 词：四轮转向 ANFIS 控制策略 仿真分析 

分 类 号：U461.91[机械工程—车辆工程]