Design and Implementation of Self-balancing and Navigation Robot Based on ROS System

摘要

Based on the SLAM (instant location and map construction) technology of monocular vision and multi-sensor information fusion, the following robot self-balancing path planning system is proposed. The system can handle strenuous moving images well, and for scenes with simple description and feature points being scarce, no loss of positioning occurs. In the selection of key frames and the connection of local maps, a variety of algorithm optimization and multi-sensor fusion methods are selected, which makes the drawing faster and more accurate. Path planning uses a more accurate and time-saving bionic algorithm-ant colony algorithm, combined with gyroscope and PID processing feedback information to achieve self-balancing and navigation path planning.