Co-ordinated Tracking and Planning Using Air and Ground Vehicles

Abstract

The MAV ’08 competition in Agra, India focused on the problem of using air and ground vehicles to locate and rescue hostages being held in a remote building. Executing this mission required addressing a number of technical challenges. The first such technical challenge was the design and operation of a micro air vehicle (MAV) capable of flying the necessary distance and carrying a sensor payload for localizing the hostages. The second technical challenge was the design and implementation of vision and state estimation algorithms to detect and track ground adversaries guarding the hostages. The third technical challenge was the design and implementation of robust planning algorithms that could co-ordinate with the MAV state estimates and generate tactical motion plans for ground vehicles to reach the hostage location without detection by the ground adversaries. In this paper we describe our solutions to these challenges. Firstly, we summarize the design of our micro air vehicle, focusing on the navigation and sensing payload. Secondly, we describe the vision and state estimation algorithms used to track ground features through a sequence of images from the MAV, including stationary obstacles and moving adversaries. Thirdly, we describe the planning algorithm used to generate motion plans to allow the ground vehicles to approach the hostage building undetected by adversaries tracked from the air. Finally, we provide results of our system’s performance during the mission execution.