Research On Tracking Control Of Dynamic Trajectories Of Mobile Robot Based On RBF Neural Network

Due to flexibility,versatility and good adaptability to the environment,wheeled mobile robots,which are a collection of many disciplines,such as machinery,computer and automatic control,have become one of the centers of robotics research,and represent the highest achievements of mechatronics.Among them,trajectory tracking control is the most important technology,and further research is still needed to improve the accuracy.Firstly,for the problem of trajectory tracking control,a typical mathematical model of wheeled mobile robot is established.Then,a kinematic controller,based on kinematics and Backstepping algorithm,is proposed,the RBF neural network is used to self-tuning the control gain in the controller to enhance the adaptability of the algorithm.In the controller,the adaptive RBF neural network and the dynamic control of neural sliding mode are introduced to solve the uncertainty of nonlinear function parameters in dynamic tracking error in the form of valuation approximation and the trajectory tracking control rate is obtained.By comparing the tracking effect with the tracking algorithm based only on sliding mode control in many uncertain cases,the experimental results show that the proposed method can effectively eliminate the trajectory tracking error of the robot and has better motion quality.Finally,through modular programming,the host computer control program is completed by MFC application framework in the VS2012,and the lower computer control program is achieved by PLC model in the TwinCAT3.Throughout the experimental operation,the actual running speed of the wheeled mobile robot is captured by the visual camera,and then the tracking control of the wheeled mobile robot is realized by compensating the speed difference,and the experiments verify the effectiveness of the algorithm.