Decision-Theoretic Multi-Agent Sensor Planning

This project investigates a decision-theoretic approach to cooperative sensor planning between multiple autonomous vehicles with specific applications for executing military missions. During the deployment of autonomous vehicles, intelligent cooperative reasoning must be used to select optimal vehicle viewing locations and select optimal camera pan and tilt angles throughout the mission. Decisions can be made in order to maximize the value of information gained by the sensors while maintaining vehicle stealth. Changes in the battlefield over time can be used to learn patterns of enemy movement and improve estimation of future utility for sensor placement alternatives. Because military missions involve multiple vehicles, cooperation can be used to balance the work load and to increase information gain. Our approach is being applied within DARPA's Unmanned Ground Vehicle program. This research is in collaboration with Dr. Diane Cook and Dr. Larry Holder from the UTA's CSE Department.