Simultaneous Localization and Human Following for a Wheelchair Robot using a Camera with Depth Information

Abstract

Abstract Powered wheelchair users often struggle to drive safely and accompanied by a caregiver. To improve assistance efficiency and reduce caregiver’s load, the wheelchair needs to autonomously follow the caregiver in the surroundings. In this paper, we present an approach to simultaneous localization and human following for a wheelchair robot equipped with Microsoft Kinect that captures RGB images along with per-pixel depth information (RGB-D camera). The speeded-up robust feature (SURF) algorithm is employed to provide the robust description of feature for environments and the target person from RGB images. Based on the environmental SURF features, we utilize an extended Kalman filter (EKF) based simultaneous localization and mapping (SLAM) to estimate robot pose. Meanwhile, a depth clustering based human detection is proposed to extract human candidates. Then target person can be recognized and tracked by combining SURF feature matching with a specified identity signature and human detection results. Moreover, a fuzzy based controller provides dynamical human following for the wheelchair robot with a desired interval. Consequently, the experimental results demonstrated the effectiveness and feasibility in real world environments.