基于神经网络的移动机器人预瞄轨迹跟踪控制算法研究

A study of control algorithm of pre-trajectory tracking for mobile robot based on neural network

轨迹跟踪控制是移动机器人运动控制中很复杂的问题。针对该问题,设计一种基于神经网络的预瞄轨迹跟踪控制算法。为了保证机器人具有较好的轨迹跟踪性能,借鉴驾驶车辆的预瞄模型,即控制驱动轮输出转矩,使得机器人在预瞄时间内到达预瞄位置。建立移动机器人动力学方程,搭建移动机器人动力学simulink模型,通过仿真获取在移动机器人不同状态下的实验数据。采用BP神经网络对实验数据进行训练,对驱动轮的驱动扭矩进行控制。最后通过仿真实验对该轨迹跟踪控制算法的有效性进行验证。

It is complicated for mobile robot to control the trajectory tracking.Thus a control algorithm of pre-trajectory tracking based on neural network is designed.A pre-trajectory driving model which means to control the driving wheel output torque,is set up in order to make the robot reach the pre-trajectory position in the pre-trajectory time.The kinetic equations as well as the simulink models are established for mobile robot.Simulation is done on the robot to obtain the in-kind experimental data.BP neural network is applied to training the data and controling the diving wheel torque.Finally the validity of this algorithm is tested through the simulation experiments.