A Networked Decision and Information System for increased agility in teaming unmanned combat vehicles

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

“…In [1], we presented one-step lookahead rollout policies π u 1,P i = {μ 1,Pi , ..., μ N −1,Pi } that optimally control the blue team formations despite the adversarial environment for the scenario at hand. The stochastic game was solved in such a way that policies for B and R play the role of minimizer and maximizer, respectively,…”

“…J in (2) represents the negative of the expected number of remaining munitions at T . The policy derived in [1] can be symbolically expressed as…”

“…Intermittent wireless communications breakdowns were not addressed in [1], despite the facts such events are bound to occur in urban environments and have the potential to adversely affect mission success if not handled correctly. In this paper, we propose to remedy this situation.…”

“…In [1], we proposed a decision policy for the routing and munitions management (RMM) of multiple UCVs evolving in an imperfectly known and adversarial environment. Intermittent wireless communications breakdowns were not addressed in [1], despite the facts such events are bound to occur in urban environments and have the potential to adversely affect mission success if not handled correctly.…”

“…Therefore, an online policy improvement step is performed during the course of the mission by utilizing the most likely GULC, as obtained from a recursive Bayesian filter (RBF). The expected cost-to-go function of the policy improvement step is approximated, by means of Monte-Carlo simulations, and the decisionmaking is then updated [1]. To carry out the online computations required to solve the RMM problem, the health state must be exchanged among the formations through the wireless network connecting the UCVs, labeled as RMM-Net.…”