车辆横向稳定性自适应神经网络控制策略研究

A Study on the Control Strategy of Adaptive Neural Network for Vehicle Yaw Stability

设计了一种"前馈+反馈"自适应神经网络控制器,通过直接横摆力矩和前轮主动转向的复合控制来提高车辆横向稳定性。反馈控制器采用PD控制策略,以实际横摆角速度与目标横摆角速度的偏差为输入量;前馈控制器采用RBF神经网络,以反馈控制器的输出为误差进行学习,而实现自适应控制。仿真结果表明,采用上述复合控制,能有效跟踪目标横摆角速度并降低质心侧偏角,提高了车辆在高速急转向时的稳定性。

An adaptive neural network controller with both feed-forward and feedback is designed, which improves the yaw stability of vehicle by combined control of direct yaw moment and active front steering. The feed- back controller adopts PD control strategy with the deviation of actual yaw rate from objective yaw rate as its input; while the feed-forward controller uses RBF neural network with the output of feedback controller as the error for learning to achieve adaptive control. The results of simulation show that using above-mentioned combined control can effectively track the objective yaw rate and reduce the sideslip angle of mass center, so enhancing the stability of vehicle in high-speed sharp turning