Research On Localization And Navigation Method Using Lidar For Medical Loading Robot

At present,a large number of human resources are still needed in the pharmaceutical process,which can not meet the requirements of aseptic,unmanned and automated in pharmaceutical process that GMP standard propose.In order to ensure the aseptic production of medical products,more and more medical loading robots with the characteristics of unmanned,aseptic and high degree of freedom appear in the pharmaceutical automation production line.In this paper,according to the current research situation,aiming at the problem of autonomous movement of indoor mobile robot in unknown environment,the localization and navigation technology of medical loading robot based on 2D lidar is studied and realized.First of all,this paper introduces the core technology of indoor mobile robot,including localization,map building,path planning and other research status,and specifically introduces its kinematics model,sensor model and so on.Then,for the indoor simultaneous localization and mapping technology,this paper studies the 2D laser SLAM algorithm,and three kinds of laser slam algorithms,Gmapping,Hector and Cartographer,are compared and analyzed,including their advantages,disadvantages and applicable scenarios.In this paper,a laser slam method based on sensor fusion is proposed based on Hector algorithm to solve the problem of location in non-flat environment and mobile robot moving rapidly.Methods the precise information of inertial navigation unit was obtained by Kalman filter,so as to determine the pose of robot in three-dimensional pose space,correct the laser data,estimate the pose of robot by IMU integration,match the frames based on Gauss Newton optimization,and avoid falling into the local optimal situation.The experimental results show that the algorithm avoids the distortion of laser data,enhances the robustness of location and improves the accuracy of map building.As for navigation technology,this paper studies global path planning algorithm and local path planning algorithm.In this paper,A~*algorithm,Dijkstra algorithm,artificial potential field algorithm and DWA algorithm are studied and implemented,and their advantages and disadvantages are analyzed through experiments.In view of the problems of more search nodes and more bends in A~*algorithm,this paper proposes a method to improve the heuristic function and increase the weight factor for optimization.The experimental results show that the algorithm can effectively reduce the search nodes,speed up the response time,and improve the real-time and smoothness of the mobile robot navigation.Finally,to prove the 2D laser SLAM method we improved can be applied to the indoor environment,the indoor environment map is constructed through experiments,and navigation experiments are carried out on this basis.The experimental results show that the system can complete the mapping and ensure that the mobile robot can reach the end point safely and stably according to the specified requirements.