Research And Implementation Of Robot Mapping Based On Multi-sensor Information Fusion

The research of robot mapping is the most basic and core technology in the field of mobile robot which develops rapidly in recent years.In order to realize accurate operation of robot in complex environment,high precision point cloud map should be taken as reference.The autonomous movement process of a robot in the environment can be divided into the following steps: The environmental data are collected by various sensors,and the original data are transmitted to the robot operating system of the industrial computer.The data are processed according to the predetermined rules,and the action instructions or attitude,position data are obtained.The action instructions are transmitted to the motor driving part,and the attitude position and other data are stored and processed in the next step.In this paper,a multi-sensor information fusion mapping method is proposed,which takes 3D solid-state lidar as the main sensor and GPS(Global Position System)and IMU(Inertial Measurement Units)as the auxiliary.On the basis of tracked robot hardware platform,built multi-sensor fusion method validation and improvement,by evaluating the effects of different fusion methods in outdoor scene,completed the robot mapping validation and improvement,to a certain extent,improved the performance of the robot to build high precision map.This paper mainly contains three parts:(1)Construction of hardware platform for tracked mobile robot.The preset application scenario of this robot platform is: autonomous patrol of outdoor fixed route.The robot platform is equipped with embedded industrial computer,3D solid-state lidar,high-precision GPS locator,low-delay high-definition image transmission module,ultrasonic obstacle avoidance sensor,IMU and so on.The 3D solid-state lidar is used as the main sensor to carry out basic map building tests.The hardware module purchased in this paper takes the complex outdoor environment into full consideration and has good performance parameters and scene applicability.(2)Theoretical elaboration of robot sensor system.Based on the mature ROS(Robot Operating System)and the Robot hardware platform built,the signals collected by various sensors are respectively sent to the industrial computer for preprocessing.The surrounding environment of the traveling route is scanned and mapped by 3D solid-state lidar to obtain 3D point cloud map.By the high precision GPS locator in accordance with the frequency of 1Hz to obtain real-time latitude and longitude and height;The acceleration and azimuth of the robot are obtained by the IMU.Through the improved sensor position layout and signal preprocessing method,the robot has better environment detection ability.(3)Theory and experimental test of multiple sensor information fusion processing.The basic theory of multi-source fusion,the classification of multi-source fusion,the concrete method and process of multi-source fusion are described.Through the outdoor road mapping test,the influence of different information fusion methods on robot mapping performance is compared.Finally,the fusion method verification of robot outdoor road mapping is completed,and the performance is improved to a certain extent.