Map-based navigation and control of mobile robots with surveillance cameras

Abstract

Recent research study on obstacle detection and path planning in unknown environments has been identified as a key issue in the development of autonomous mobile robotic systems. This paper presents a path planning and control strategy for blind robots in indoor environments. In particular, a systematic approach is employed to convert 3D and 2D navigation maps from a 3D sparse map established by a scout robot equipped with onboard vision using a VSLAM (visual simultaneous localisation and mapping) algorithm. Based on the navigation map, desired trajectory can be planned for a blind robot to reach the target. An effective image-based control law is proposed to drive the blind robot to follow the planned trajectory with real-time monocular camera vision. The proposed navigation and control system has been validated by simulations and experiments with a blind robot using a 3D sparse map and real-time vision from surveillance cameras.