Design And Research On Mapping And Navigation For Mobile Robot Based On Multisensor Fusion

With the development of technology and economic,robots have gradually become an integral part of production and life.Especially in recent years,with the rise of robots combined with artificial intelligence,human-robot collaboration has become the trend of research in the field of robotics.The mobile underpan equipped with human-robot collaboration robot arm has gradually become the focus of research in the field of mobile robots.This project takes the mobile robot as its application background,carrying out the exploration of robot technology in localization and mapping,navigation and path planning.Also,a mobile robot system integrating mechanical hardware,embedded software application,mapping and path planning are studied and built.And multi-sensor fusion technology is applied to mobile robot,improving the accuracy of localization and the efficiency of path planning.Firstly,a comprehensive framework for mobile robot that can autonomously localize and navigate is researched and built.The mobile robot can make a turn in zero radius,so that move flexibly indoors.An independent suspension device is designed to improve the movement of the robot on uneven surfaces.And scalable motor servo control system is constructed,providing interface for the control of upper robot arm system.The external environment through Lidar and depth camera is detected,and MCU is used to integrate various safety devices,such as anti-collision rubber strip,ultrasonic sensor,anti-drop sensor,to ensure the safety of the mobile robot when it is running.ROS is taken as the software system,which can increase the efficiency of program operation and data transmission and ens ure the stability of the system.Then the motion control and pose estimation for the two-wheeled differential drive mobile robot are studied,and the odometer model and the motion control model are established.Taking the advantages of IMU to compensate for accumulated odometer error caused by wheel slippage and environmental changes of the encoder,such as high accuracy in a short time and rapid dynamic response.Extended Kalman filter is used to combine the encoder and IMU,so that the mobile robot's pose estimation model is constructed and the accuracy of the Lidar SLAM algorithm is improved.Moreover,aiming at the defect of the SLAM can't detect the obstacles in the three-dimensional environment in a complex environment only with Lidar,Bayesian formula is applied to combine Lidar and depth camera,so as to create an indoor Grid map for 3D environment.Making full use of the redundant information of laser and vision sensor data to get an effective indoor map.Making full use of the redundant information of laser and vision sensors data to get an effective indoor map.According to the moving obstacles detected by the Lidar and depth camera,the speed field of obstacles is added to the potential field method.And a dynamic obstacle avoidance model based on the improved potential field method is constructed,which improves the local path planning efficiency of the mobile robot.Finally,a two-wheeled differential mobile robot experimental platform is built to complete the integration of IMU's odometer SLAM experiment,building the three-dimensional grid map with the integration of Lidar and depth camera,and completing the dynamic obstacle avoidance experiment based on the improve d potential field method,so as to verify the feasibility of the algorithms.