A dual-network health state estimator and decision policy for unmanned combat teams

We propose a one-step lookahead rollout policy in closed-loop with a health state estimator to ensure effective cooperation among unmanned combat teams despite intermittent wireless communications breakdowns. To ensure effective cooperation despite network faults, the proposed scheme relies on dual networks. On the one hand, a sensory information management network (SIM-Net) provides the most probable distribution on the location and classification of the adversarial ground units by fusing mobile sensor measurements obtained by a team of surveillance vehicles. On the other hand, a routing and munitions management network (RMM-Net) enables unmanned combat vehicle (UCV) communications, which are required for their effective path planning and for the distribution of the rollout decision policy over the formations. Simulation results demonstrate the effectiveness of the proposed health state estimator and decision policy.