High-precision docking of wheelchair/beds through LIDAR and visual information

Abstract

To address the low docking accuracy of existing robotic wheelchair/beds, this study proposes an automatic docking framework integrating light detection and ranging (LIDAR), visual positioning, and laser ranging. First, a mobile chassis was designed for an intelligent wheelchair/bed with independent four-wheel steering. In the remote guidance phase, the simultaneous localization and mapping (SLAM) algorithm was employed to construct an environment map, achieving remote guidance and obstacle avoidance through the integration of LIDAR, inertial measurement unit (IMU), and an improved A* algorithm. In the mid-range pose determination and positioning phase, the IMU module and vision system on the wheelchair/bed collected coordinate and path information marked by quick response (QR) code labels to adjust the relative pose between the wheelchair/bed and bed frame. Finally, in the short-range precise docking phase, laser triangulation ranging was utilized to achieve precise automatic docking between the wheelchair/bed and the bed frame. The results of multiple experiments show that the proposed method significantly improves the docking accuracy of the intelligent wheelchair/bed.