Mapping And Localization Method Of Inspection Robot Based On SLAM And Its Iterative Learning Control

With the rapid development of artificial intelligence in recent years,inspection robots have begun to be increasingly used in service,military and other industries.There are different types of inspection robots for different application scenarios.This article mainly studies an outdoor inspec-tion robot,and researches its inspection plan,system design,control methods,and other issues,with the aim of giving a more complete inspection robot design plan.This paper first introduces and verifies a GPS-based inspection plan.In view of the limitations of this scheme,based on simultaneous mapping and localization(SLAM),the research and proposes the construction and positioning methods of inspection robots.The patrol control method based on iterative learning control is further proposed.The specific research contents include:1.Relying on a cooperation project with an institute of the Chinese Academy of Sciences,an electric inspection vehicle was used as a test platform to complete a GPS-based inspection system.Through experiments,the effectiveness of the scheme in open scenes was verified;2.Aiming at the shortcomings of the GPS inspection system being vulnerable to building oc-clusion and unstable signals,a submap-based mapping algorithm was designed and built to add closed-loop detection and closed-loop optimization functions to the laser odometer.In addition,multi-thread acceleration is used to improve the point cloud matching speed.The experimental results show that the iterative mapping algorithm can construct a high-precision point cloud map;3.An IMU-based loosely coupled positioning system is proposed,which uses pre-integration technology to predict poses in real time,speeds up algorithm convergence,and does not re-quire kinematic models.In the back-end sliding window optimization,IMU pre-integration residual term and prior pose constraint are added to estimate the IMU bias.Experiments show that the loosely coupled system can achieve accurate pose estimation and improve relocation accuracy;4.Improve on the basis of the original iterative learning control method of quadratic criterion Most of the current methods only use position information.In this regard,a speed indicator is added to the objective function to derive the closed loop of the improved iteration learning control law of the secondary criterion Formula,and theoretical analysis of the convergence of the algorithm.In addition,in order to solve the problems of uncertain parameters of the robot system,state estimation and parameter estimation methods are further introduced to improve the control performance.The convergence performance of the improved quadratic criterion iterative learning control is verified through experiments,and it is applied to the tracking control of inspection robots.