Research On Path Planning Of Indoor Mobile Robot Based On 2D-SLAM

With the continuous progress of science and technology and the continuous improvement of living standards,robots have become an indispensable part of human life.The most common of them are mobile robots,which are used to provide services for people,such as food delivery robots,sweeping robots,and hospital robots.Medicine delivery robot,etc.As the most important technologies in mobile robots,Simultaneous Localization and Mapping^[1]（SLAM）and path planning technology have been developed and researched for many years.In this study,a four-wheeled mobile robot with Lidar as a sensor was used to conduct theoretical research and simulation experiments on typical 2D-SLAM algorithms and path planning techniques.A Robot Operation System（ROS）was used to design a A low-cost,high-precision four-wheel mobile robot that can complete high-precision mapping and path planning.Firstly,the coordinate system establishment,kinematic analysis,sensor modeling and map type analysis of the four-wheeled mobile robot used in this paper are carried out,and the classical ranging models are compared,and the likelihood field model is selected as the ranging model in this paper.Three commonly used map types are selected,and raster map is selected as the map type for subsequent experiments.Then,the principle of Extended Kalman Filter SLAM is mainly studied,and then the SLAM algorithm of the Improved Extended Kalman Filter（Improved Extended Kalman Filter）proposed in this paper is theoretically derived.Based on the superiority of the original algorithm,the classical 2D-SLAM algorithm,Hector and Gmapping algorithms are theoretically studied,and the theoretical estimation comparison chart and error analysis of the three algorithms are carried out,and the efficiency of the improved algorithm is determined.Secondly,the global path planning algorithm and the typical algorithm in the local path planning algorithm are studied respectively,the A*algorithm and the D*algorithm in the global path planning algorithm are theoretically studied and simulated experiments are carried out,and then the local path planning algorithm is studied.Dynamic window method（Dynamic Window Approach,DWA）^[2]was deduced and simulated experiments and analyzed the results.Finally,the operation rules of traditional ant colony algorithm were analyzed,and it was easy to fall into local optimum,long running time,Due to the problem of too many iterations,an improved ant colony algorithm is proposed,and simulation experiments are carried out for the two,and several sets of experiments are compared to prove the superiority of the improved algorithm.Finally,the construction of the hardware and software platform of the four-wheeled mobile robot is introduced,and the completed four-wheeled mobile robot is used as the experimental platform to conduct mapping and path planning experiments.The SLAM algorithm based on the improved extended Kalman filter proposed in this paper is transplanted to the four-wheeled mobile robot platform,and compared with the Hector and Gmapping algorithms,which proves the superiority of the algorithm proposed in this paper.Finally,use the constructed map to conduct path planning experiments,observe the driving situation of the mobile robot through visualization tools,and analyze the experimental results.