基于神经网络的机器人驱动系统数学模型辨识

Mathematical Model Identification of Robot Drive System Based on Neural Network

以轮式移动机器人驱动系统为主要研究对象,提出一种基于神经网络的分析与试验相结合的建模方法。该方法在分析理想直流电动机数学模型的基础上,采用串-并行辨识结构的神经网络确定系统输出与输入的关系,实现了机器人驱动系统的数学模型辨识。最后,通过试验证明了该方法的有效性,为未来工、农业等领域机器人的精确运动和动力控制的研究提供了技术支持。

This paper studies mainly a wheeled mobile robot drive system with modeling method proposed based on combination of neural network analysis and experiment. Through the modeling method which is based on analysis of the ideal DC motor mathematic mode,, the mathematical model identification of robot drive system has been achieved by using neural network with serial-parallel identification to coilfirm the relationship between the input and output of the system. This method has been proved available through experiment, and will be the technical support of the future research of accurate motion and drive control in the industrial and agricultural field,etc.