Modeling And Path Tracking Control System Of Spherical Robot With Visual Technology

Spherical robot is a kind of robot containing various devices inside the spherical shell to realize rolling autonomous movement.It has unique advantages and wide application prospects in the fields of life entertainment,industry,military and space exploration.In order to improve the environmental sensing ability and control performance of the spherical robot,this thesis designs and implements a spherical robot with a vision device separated from the sphere,and studies its mechanical structure design,motion performance,data-driven modeling method and path tracking control system.Firstly,a spherical robot based on visual feedback is designed.The visual device and the sphere are designed independently to obtain the external environment information without shelter.To improve the accuracy of the posture and motion parameters of the spherical robot,a cloud head was built to connect the visual device through magnetic force.At the same time,the driving cart composed of three omni-directional wheels with 120° distribution to realize omni-directional motion of the ball.After elaborating the mechanical structure design of the spherical robot,the hardware and software design of the system are introduced.Secondly,a modeling method for spherical robot based on General Regression Neural Network(GRNN)is proposed.On the basis of the kinematic performance analysis of the spherical robot,aiming at its nonlinear,multi-variable,strong coupling,parameter uncertainty and other complex characteristics,it is difficult to establish an accurate model,this thesis combines GRNN and improved gray Wolf algorithm to establish a data-driven spherical robot model.Simulation experiments verify the feasibility and effectiveness of the modeling method.Thirdly,a spherical robot path tracking system based on Sliding Mode Control(SMC)and GRNN supervision control is designed.In order to overcome the problem that the spherical robot is susceptible to perturbation,a SMC feedback controller of the spherical robot is designed.The GRNN feedforward controller generates the teacher’s signal to learn the SMC output,and forms the feedforward + feedback position supervision and control system of the spherical robot to realize the stable path tracking of the spherical robot.Simulation results show that this control strategy can improve the path tracking accuracy and adaptive ability of the system.Finally,the path tracking experiment platform of the spherical robot was built,and the path tracking experiment of the spherical robot was completed.The results show that the spherical robot system designed in this thesis can track the set path effectively.