Design And Implementation Of Control System Of Mobile Robot Based On Multi-sensor Fusion SLAM Framework

With the continuous development of science and technology,robot technology has made breakthrough progress,more and more automatic robots are applied to all aspects of production and life.As an important branch of robot technology,mobile robot technology is also one of the current hotspots of academic research.The core technology of mobile robots is SLAM(simultaneous localization and mapping).The mainstream research directions are LiDAR-SLAM and Visual-SLAM.These two SLAM technologies have their own advantages and disadvantages.Based on the factory indoor environment as the application scene,this paper studies the SLAM framework and algorithm of multi-sensor fusion using LiDAR and vision sensors.The main contents are as follows:1.Based on the grid map and point cloud map alignment matching algorithm,this paper aligns the grid map generated by LiDAR-SLAM with the point cloud map generated by Visual-SLAM through the custom algorithm,which can improve the accuracy of global positioning in the highly repetitive factory environment.2.A new SLAM framework based on low-cost LiDAR and vision is proposed,which optimizes the robot's posture by combining the laser scanning matching results and image feature points data as the constraints of graph optimization method.Visual features are also used for closed-loop detection,to optimize the grid map constructed by the SLAM framework.3.this paper designs a new factory instrument monitoring robot management control system,which consists of three parts:mobile robot,server,and mobile client.The mobile robot can achieve a series of functions such as image acquisition,user management,displaying trajectory of the robot and mapping in a factory indoor environment.