Research On Autonomous Navigation And Pedestrian Tracking Of Indoor Mobile Robot Based On Laser Lidar

Since the 21 st century,artificial intelligence has gradually taken the forefront of the science and technology stage,and the research on robotics gradually becomes more and more intelligent.Indoor mobile robots are an important research direction of intelligent robots.Autonomous navigation and pedestrian detection tracking are important functions of indoor mobile robots which insure robots to carry out other tasks.It is of great importance to study autonomous navigation and pedestrian detection tracking of indoor mobile robots.Based on the application background of indoor mobile robot,the overall design of indoor mobile robot is given in this paper firstly.Through the description of SLAM problem and the analization of the common SLAM problem solving method,the gmapping algorithm is proposed as the location and mapping method.The parameters of scanning matching score and the number of particles are simulated in order to adapt to the environment of this paper.Secondly,the autonomous navigation technology of indoor mobile robot is researched.By comparing the advantages and disadvantages of many robot path planning algorithms,the heuristic global path planning algorithm is analyzed.According to the advantages and disadvantages of traditional Dijkstra and A * algorithms,an A* algorithm with improved valuation method is proposed as the global path planning algorithm and is simulated in MATLAB in order to improve the shortcoming of many turning points of traditional A * algorithm planning.According to the characteristics of many obstacles in the complex indoor environment,a local path planning method based on DWA dynamic window algorithm is proposed,and is simulated in MATLAB to select the parameters of the evaluation function.Thirdly,the pedestrian detection tracking based on Lidar is studied in this paper.A method of data clustering and environment segmentation based on hierarchical clustering algorithm is proposed and is simulated to determine the best parameters.According to the lack of information based on lidar 2D data,this paper proposes a pedestrian detection feature based on double-legged pedestrian model to realize pedestrian detection in indoor complex environment.The Kalman filter is used to track the detect results and a method to lock the tracking target is given to realize the robot's automatic follow-up pedestrians.Finally,the methods proposed in this paper are implemented in the ROS robot operating system.Experiments are carried out using the choosing robot platform to verify the feasibility of the proposed methods.The results of the experiments show that the robot achieves autonomous navigation and obstacle avoidance in indoor environment and realizes the detection and tracking of pedestrians which shows that the studies in this paper has certain practical value.