Autonomous Planning and Replanning for Mine-Sweeping Unmanned Underwater Vehicles
- Wednesday, 01 September 2010
This software generates high-quality plans for carrying out mine-sweeping activities under resource constraints. The autonomous planning and replanning system for unmanned underwater vehicles (UUVs) takes as input a set of prioritized mine-sweep regions, and a specification of available UUV resources including available battery energy, data storage, and time available for accomplishing the mission. Mine-sweep areas vary in location, size of area to be swept, and importance of the region. The planner also works with a model of the UUV, as well as a model of the power consumption of the vehicle when idle and when moving.
The planner begins by using a depth-first, branch-and-bound search algorithm to find an optimal mine sweep to maximize the value of the mine-sweep regions included in the plan, subjected to available resources. The software issues task commands to an underlying control architecture to carry out the activities on the vehicle, and to receive updates on the state of the world and the vehicle. During plan execution, the planner uses updates from the control system to make updates to the predictions of the vehicle and world states. The effects of these updates are propagated into the future and allow the planner to detect conflicts ahead of time, or to identify any resource surplus that might exist and could allow the planner to include additional mine-sweep regions.