Design of Robot Positioning System Based on the Integration of Slam System and Inertial System

Abstract

Slam robot positioning system based on vision has good universality, but the depth information of the environment is lost due to low bandwidth, uncertainty of visual image change and poor real-time performance of motion mutation. In order to solve the above problems, a robot positioning system based on the integration of slam system and the inertial system is proposed. The hardware part designs inertial sensor modules such as accelerometer, odometer and gyroscope, and integrates the inertial sensor module with slam system. The IMU motion model is used to predict the attitude of the camera in the current frame and match the feature points of the current frame. The state of the robot is estimated by using the inertial sensor data of binocular vision, and the position and attitude of the robot are determined by slam according to the estimated value. The system test results show that the maximum positioning error rate of the system is only 1.8%, the positioning time is short and the performance is good.