Outdoor Navigation of a Mobile Robot by Following GPS Waypoints and Local Pedestrian Lane

摘要

Combining global GPS-based navigation and local pedestrian lane following is an promising way to autonomous navigation of mobile robots in unknown outdoor environments. By taking advantage of open SDK online satellite maps, the navigation route which contains a sequence of intersection waypoints could be autonomously planned. Through moving toward the GPS coordinates of the waypoints sequentially, the robot could be piloted to the destination in a global and coarse-grained navigation manner. Meanwhile, a simple vision-based lane following method is proposed to control the robot to move safely on the pedestrian lane in a local and fine-grained navigation manner. The two types of controllers are implemented as two behaviors which are managed by a Finite State Automaton (FSA)and coordinated with an obstacle avoiding behavior. The overall navigation system is implemented on the HunterBot platform with a cloud based human-robot interaction system. Elementary blind test results on the campus environments have shown that the proposed method is potentially effective to address the autonomous navigation problem in unknown outdoor environments.