Information Exchange and Decision Making in Micro Aerial Vehicle Networks for Cooperative Search

This paper considers a network of autonomous micro aerial vehicles (MAVs) cooperatively searching for multiple stationary targets. The objective is to minimize the search time while considering sensing and communication limitations. We derive limits for the required number of sensor observations considering false alarms and misdetections, to declare the existence or absence of a target. To minimize the search time, we formulate the cooperative search as a traveling salesman problem and use the calculated sensor observation's limits in each predicted location of the target. Moreover, the design options in multi-MAV cooperative search are explored and classified in two dimensions: information merging and decision making where each dimension can be either centralized or distributed. Algorithms are then introduced to analyze the effects of centralized or distributed coordination for minimizing the search time. We show that depending on the availability of information and capability of making decisions, the MAVs can search an area more efficiently if information merging and decision making are distributed. We compare our proposed algorithms with a traditional lawnmower search pattern and a distributed cooperative search algorithm.