大型轮式工程车辆转向系统的神经网络PID控制

Research on Neural Network PID Control with Application to Heavy-duty Wheeled Vehicle Steering System

根据大型轮式工程车辆转向系统的对象特点和操纵方式,提出采用基于RBF神经网络控制器来改进常规PID控制器实现系统控制性能。该控制系统结构中,RBF神经网络辨识器(RBFNNI)实现对被控对象的Jacobian矩阵信息的辨识,神经网络控制器(NNC)是基于RBF神经网络实现的单神经元的PID控制器。在对算法进行改进的基础上设计了神经网络结构,并进行了被控对象的仿真分析。实际结果表明该控制方法具有较好的实用性和鲁棒性,可以用于多操纵模式工程车辆转向系统的控制。

According to the plant characteristics and the operating method of the heavy-duty wheeled vehicle, a neural network controller is proposed for improving the control performance of the traditional PID controller. The configuration of the control system is based on RBF neural networks, one is RBF neural network identifier (RBFNNI) used to identify the Jacobian matrix of control plant, and the other is neural network controller (NNC) used to provide nonlinear PID control parameters, which is used to achieve the better performance than conventional single PID controller. Then, neural network architecture is designed according to improved control algorithm. At last, simulation result is given and discussed. The result shows that the control system is robust and adaptive in dealing with nonlinear systems, so it is feasible for control steering and manipulating system of the wheeled vehicle.