基于Simulink的四轮转向汽车神经网络控制策略仿真

针对汽车小转角时质心侧偏角为零,高速大转角时前轴抗侧滑的控制目标,提出一种四轮转向汽车控制策略.在Simulink环境下建立包含轮胎非线性和计及侧倾的三自由度四轮转向汽车模型,运用双隐含层BP神经网络训练得到四轮转向控制器.仿真结果表明,神经网络控制器可有效控制高速时汽车前轴滑动的趋势,并在低速到高速时使汽车质心侧偏角基本为零,控制误差低于比例转角控制策略和横摆角速度反馈控制策略.同时高速时横摆角速度响应与前轮转向汽车接近,汽车的侧向加速度和车身侧倾角稳态值比前轮转向有所降低.

Simulation of neural network control strategy for four-wheel-steering vehicle based on Simulink

Aimed at keeping side slip angle at zero while inputting small steering angle,and keeping front axle anti-slip while inputting large steering angle,a control strategy for four-wheel steering(4WS) vehicle was proposed.Three degrees of freedom 4WS vehicle dynamic model including nonlinear characteristic of tyres and roll motion was established with Simulink.The neural network controller of 4WS vehicle was constructed based on double hidden layers BP neural network.Simulation results show that the neural network controller keeps front axle from side slip and maintains side slip angle at zero in most cases,and the control errors are less than those in the proportional angle control and the yaw rate feedback control.In addition,the amplitude of yaw rate is similar to that of front-wheel-steering vehicle and the steady response of lateral acceleration and the body roll angle are decreased than front wheel steering vehicle.