Research on Sensor Technology in Mobile Robot Navigation

As mobile robots are widely used in daily life, industrial manufacturing and the military, their ability to autonomous navigation in unmanned platforms and a wide range of environments is increasingly demanding. Therefore, the selection of sensors is a necessary process to improve the efficiency of navigation. The paper will introduce the principles and advantages of monocular vision, LIDAR and ultrasonic sensors etc. in detail, and then explore the advantages and disadvantages of various algorithms, and finally conclude optimal fusion of sensor solutions. The comparison results present that for the monocular vision, acquiring an image from a single camera setup, then using a YOLOv5 and mosaic technology to form a single image and finally using the improved RRT and the Frenet coordinate system to model the path is an efficient solution. 3D LIDAR technology can use the SLAM framework of graph optimization to create the map for obstacle avoidance and path planning. At last, this paper provides suggestions and optimizations for mobile robot navigation solutions, which are integrating multiple sensors and combining navigation solutions with machine learning.