Cooperative Multi-UAV Active SLAM for Traversing Partially Mapped Terrain

This paper develops both single and multi-UAV path planning schemes for navigating over partially known terrain in GPS-denied environments. The UAVs perform Simultaneous Localisation And Mapping (SLAM) using a combination of on-board inertial sensors and an on-board terrain sensor and share map information with one another over a data fusion network. UAV paths across the terrain are planned such that the expected SLAM localisation accuracy for each vehicle is maintained above specified constraints. The path planner thus balances between traversing known and unknown sections of the terrain while attempting to reach a predefined objective point in minimum time. Simulation results are presented using a team of UAVs.