220 / 2023-09-27 00:20:30

Multi-source fusion positioning algorithm of ROS robot platform based on improved LVI-SAM

Underground complex environment；,improved LVI-SAM,multi-source fusion

The rapid acquisition of surrounding environmental information for the carrier is crucial for achieving accurate and robust positioning in underground spaces. This study focuses on optimizing the fusion of lidar, vision, and inertial navigation using the LVI-SAM algorithm to achieve robust positioning of the ROS robot platform in underground spaces. The proposed method enhances visual initialization by utilizing imu node data prediction, improves visual depth estimation with laser data, enhances the interaction of node data information by providing bias initial estimates for the imu through vision, and constructs a closed-loop factor using the global pose map to facilitate algorithm optimization. Experimental results demonstrate that the optimized algorithm effectively reduces positioning translation errors and enables high-precision and robust acquisition of position information in the underground complex field environment for the ROS robot platform.