Research On Inspection Robot Under EMUs

In order to improve the maintenance efficiency of equipment under EMUs and reduce the work intensity of the maintenance personnel,according to current technical requirements for maintenance of EMUs,a method of using inspection robots to replace the personnel for maintenance is proposed.First of all,this thesis introduces the development status of EMUs and the significance of research on inspection robots.It also analyzes and expounds the development of robot industry and technology,image recognition theory at home and abroad.In addition,this thesis analyzes and summarizes the maintenance procedures and under-vehicle maintenance items implemented by EMUs at this stage.Based on the understanding of maintenance needs on EMUs and the development status of robots,the functional requirements of the inspection robot are analyzed.Then system components are proposed,which includes manipulator,mobile delivery platform,control and processing center.A specific kind of manipulator is selected as the actuator for maintenance operation.Degrees of freedom,kinematic configuration,working space and other contents are analyzed,detailed technical parameters are defined.In addition,this thesis also analyzes and selects driving and steering method of the mobile delivery platform,the hardware composition of vision system,communication module,control and processing center.In this thesis,manipulator of the inspection robot is taken as research object,and the kinematic model of the manipulator is established in robotics D-H method.The forward and inverse solutions of the kinematic model are solved.Finally,the MATLAB Robotics Toolbox is used for simulation to verify the correctness of the forward and inverse solution.This provides a theoretical basis for motion control and trajectory planning of the manipulator.After analyzing the motion of the manipulator,the thesis plans the inspection trajectory and gives the position schedule.In addition,trajectory planning is performed on the joint space of the manipulator.MATLAB software is used to simulate the trajectory planning,which verifies the feasibility.After completing the position control for image collection,the thesis conducts automatic recognition research on these images.This thesis introduces the fault diagnosis requirements of the inspection robot,and selects artificial neural network as the image recognition method.The principles of BP neural network and convolutional neural network that are involved in the method are introduced.On this basis,an image recognition system based on artificial neural network is established for automatic diagnosis of under-vehicle faults.Taking the loose bolt fault for test,the system obtains a good recognition result.The feasibility of applying automatic fault diagnosis system,which is established by artificial neural network,to the inspection robot under EMUs is proved.At the end of the paper,it summarizes the results and proposes future research direction of the subject.