UAV-Guided Navigation of an Unmanned Ground Vehicle
(Room Innovation Hub-- Booth 2727)
03 May 16
10:00 AM
-
5:30 PM
Tracks:
Air, Ground, Research and Development
Unmanned ground vehicles (UGVs) offer many advantages for various types of mission including search and rescue and ordnance disposal. They significantly reduce the threat to human. These vehicles equipped with vision system and robotic arms can be autonomously routed to identify victims, targets, and threats and for ordnance disposal. However, their operation in urban environment and terrains poses unique challenges such as visibility constraints posed by high-rise buildings and unreliable GPS data in urban environments or in the terrains. This makes autonomous navigation of these vehicles difficult. Also, since they operate in the ground, the field of view they provide is limited. Therefor unamend aerial vehicle (UAV)-guided navigation of these vehicles is becoming increasingly important as the UAVs provide better and exocentric views. A typical scenario is for search and rescue missions where a UAV equipped with an onboard vision system identifies a victim on the ground and determines the victim’s location and provides the UGV with the location of the target. The UAV then guides the UGV to the area while continuously monitoring and sharing the environment with the UGV and increasing its situational awareness. The UGV provides on-ground evaluation and assessment and delivers a rescue package to the victim. This paper presents the research being done at Cal Poly Pomona on the UAV-guided navigation of a UGV for search and rescue type missions. In our research, a UGV is provided with a series of waypoints or coordinates to follow based on the target location identified by the UAV based vision system. The USV flies a predefined flight path autonomously using an autopilot. The path is determined by the mission boundary. The main problem of the UGV is to follow the waypoints, find the path and keep itself on path. In order to follow the waypoints, the correct robot pose, odometry, and orientation has to be estimated. A camera onboard the UGV constantly takes images to analyze the road condition. The UGV is also equipped with a laser scanner and sonars for obstacle detection. The vision module classifies images into drivable and non-drivable regions and thus identifies the path to follow. The paper will show the simulation and experimental results.