Research On The Control System Of Smart Wall-climbing Robot

With the development of industrial automation,the robot is playing an increasingly important role.Especially in complex industrial environment(such as environments with high pressure,high temperature,high radiation,etc.),robots are required for the inspection,maintenance,material transportation,and so on.The traditional industry needs to be adjusted and upgraded,so it is urgent to improve the automation level of robots' work to reduce risks of specific work and improve efficiency.With the experiment platform of the ship cleaning robot,an intelligent wall-climbing robot control system is developed in this paper based on reinforcement learning algorithm.This system is featured by openness,universality,easy operation,and high-level automation,meeting the requirements of traditional industries' upgrading.This paper mainly includes the following contents:I.Hardware design for intelligent wall-climbing robot control system based on STM32F407Based on requirements for the design of wall-climbing robot control system,an intelligent wall-climbing robot control system-two-level distribution control for upper and lower monitors has been developed in this paper.And through comparative analysis,the hardware structure is built based on STM32F407.The hardware structure includes STM32F407 and its peripheral circuits,matrix keyboard of upper monitor,communication modules and the external SRAM of lower monitor,the status display module,sensor module,communication module,servo motor drives,and motor actuator.The PCB for control circuit board of the lower monitor is also designed according to PCB principles.II.Self-tracking strategy for intelligent wall-climbing robot control system based on the tracking Q(?)algorithmFor the automatic cleaning mode of wall-climbing robots which are used for ship cleaning,the self-tracking strategy for robots is developed based on the reinforcement learning algorithm.Obtain the cleaning target track through Alpha Blending and develop the tracking Q(?)algorithm based on the tracking algorithm and Q-learning algorithm,enabling wall-climbing robots to study automatically as well as clean and track the target track,so as to solve the self-track issue for wall-climbing robots under unknown environment,simulate work scenarios,and build a mathematical model.The simulation results prove that with the preceding algorithm,the function of automatic cleaning can be realized for wall-climbing robots.III.Software design for the intelligent wall-climbing robot control systemUnder the circumstance of Keil ?Vision4 development,using C language to design the software for upper monitor,lower monitor,and RS485 communication module,and test the software modules of the intelligent wall-climbing robot control system.IV.Experiments for overall control system with wall-climbing robots which are used for ship cleaningIn the overall experiments,experiments are conducted on the manual mode and automatic mode of the control system by using experimental platform for intelligent wall-climbing robot control system—wall-climbing robots which are used for ship cleaning and experimental results are analyzed.Experimental results show that: for the control system of intelligent wall-climbing robot designed in this paper,communication signal between the upper monitor and lower monitor can control robot movements.While MATLAB simulation results show that: for the self tracking,the convergence of tracking Q(?)algorithm is more efficient than conventional reinforcement learning algorithm.Therefore,this design is developed to increase the automation level of wall-climbing robots,making robot control system become more open and easy to operate,so as to meet expected design requirements.