Cooperative reconnaissance for stationary and moving targets using mixed integer linear programming

This work focuses on the use of cooperative multi-vehicle systems in reconnaissance. A team of vehicles travels to collect information of targets such that their uncertainties are reduced. An approach is developed and employed to effectively plan paths of the vehicles. The high-level of a hierarchical control structure is considered. In prior works, the path optimization is generally constructed so that it can be used at any circumstances. While it offers many benefits, it is computationally intensive. It is proposed here to decompose the planning into 1)strategy planning and 2)task/path planning. Unlike other works, the two elements are separated. The paths are planned under the policy given by the strategy planning. The separation allows the policy to be adjusted based on the current environmental status. Therefore, the planner becomes more flexible and is able to tackle more complex situations such as moving targets and obstacles. The knowledge from an optimal planner is employed to provide an insight on the best policy. The simulations show the vehicles successfully observe multiple moving and stationary targets using relatively short run time.