Distributed and cooperative decision making for multi-UAV systems with applications to collaborative electronic warfare

In this paper, a theoretical control framework for distributed cooperative decision making for an ensemble of UAVs is presented. A decentralized optimization problem was proposed for the control of multiple UAVs for cooperative decoy jamming of an ISAR radar system. Each vehicle was treated as a socially capable and intelligent agent that used neighboring information to influence its own behavior and achieve the desired group behavior. The optimization problem was formulated using a linear superposition of individual and group tasks and decentralized using a neighborhood construction. Neighboring vehicles were coupled using the aggregate cost function and consensus protocols. To achieve coordination and consensus a decentralized control scheme based on a decentralized model predictive control strategy was developed and implemented. The decentralized model predictive control algorithm synchronized the input of neighboring vehicles and achieved a consensus in the output at each prediction horizon. The consensus of neighboring vehicles was described by the group task. The algorithm was simulated for an ensemble of small-scale UAVs for decoy jamming. Given a specific radar cross section for each UAV, the UAVs were able to self-organize and coordinate their behavior using the developed algorithm and achieve a collective radar signature comparable to a given aircraft model. Preliminary results also demonstrated optimal and closed-loop stable behavior.