Research On Indoor Mobile Robot Navigation Technology Based On Laser Vision Fusion

With the development of social productivity and economic level,the demand for robotic products is increasing,which makes the various robotics technologies developed.At the same time,strict environmental constraints,specific path requirements and higher autonomy requirements place higher demands on autonomous navigation of mobile robots.To improve the localization accuracy and strong robustness of mobile robots in complex environments to meet higher application requirements.The paper designed a laser vision fusion-based SLAM(Simultaneous Localization and Mapping)and navigation system for indoor mobile robots in complex indoor application environments and structured scenes with a single environment of features.The paper also studied the algorithms to improve the positioning accuracy and system robustness of mobile robots.First,in order to validate the reliability and robustness of the algorithm in subsequent articles,the paper designed and built the hardware system of indoor mobile robot.According to the needs of the actual application environment,the article analyzed the kinematics of the adopted two-wheel differential speed model.The article also carried out software and hardware design for mobile robot chassis motion control,and finally built a complete software and hardware platform for indoor mobile robot SLAM and navigation.Secondly,in order to better fuse the environmental sensing data from the sensors,2D Li DAR and depth camera sensor mathematical models were established,and 2D Li DAR,depth camera,IMU,and wheeled odometer were calibrated.For the external calibration of2 D Li DAR and depth camera,this paper proposed a simple and novel calibration method and calibration board for the joint calibration of 2D Li DAR and depth camera.The method greatly simplified the calibration process of both sensors and improved the accuracy of the joint calibration.Then,the multi-sensor positional estimation fusion algorithm,SLAM algorithm and path planning algorithm for indoor mobile robots were investigated.The article proposed an adaptive unscented Kalman filter(AUKF)based multi-sensor position estimation fusion algorithm for indoor mobile robots.The method improved the positioning accuracy of mobile robots while automatically adjusting the fusion weights of each sensor according to the harshness of the environment to improve the adaptability and robustness of robot positioning in complex environments.Based on the Cartographer algorithm,the paper investigated the fusion of 2D point cloud from 2D Li DAR and 3D point cloud from depth camera.The algorithm established a two-dimensional raster map with richer environmental features to improve the accuracy of sensor feature matching and the safety of robot work.Based on the A* algorithm and the artificial potential field algorithm,the article proposed an optimized path planning algorithm with the fusion of A* and artificial potential field method.The algorithm ensured optimal path planning in static environments and fast sensing and path planning in local dynamic environments,improving the reliability of robot path planning and the safety of the robot itself.Finally,the above algorithms were validated based on the built indoor mobile robot SLAM and navigation software and hardware platform and indoor experimental environment.The experimental results have shown that the proposed laser vision external reference joint calibration algorithm improved the external reference calibration accuracy.The adaptive traceless Kalman filter-based fusion algorithm improved localization accuracy of mobile robots and positioning accuracy in complex environments.The adopted SLAM algorithm and path planning algorithm enabled the mobile robot to navigate accurately and safely based on the established complete and rich raster map to meet the practical application requirements.