Cooperative and Deceptive Planning of Multiformations of Networked UCAVs in Adversarial Urban Environments
Nicolas Lechevin, C.A. Rabbath, Marc Lauzon · AIAA Guidance, Navigation, and Control Conference and Exhibit · 2007
We present an online decision policy built upon Markov decision processes for the cooperative path planning and weapons management of multiformations of UCAVs. Such cooperative control strategy provides optimal routing and weapons management despite conflicting objectives of opposing teams. The UCAVs constitute the blue team. They have for objective to reach prescribed tactical target locations, sequentially, from a common starting point, by following possibly different paths across an adversarial urban environment, within a prescribed time window and with maximum destruction capability once at close range. The UCAVs face an adversarial red team, which is composed of static ground units that can engage any nearby UCAV. The blue team’s planning thus aims at minimizing damages while maximizing the total number of remaining weapons at the time the tactical targets are reached. The blue team is modeled as controlled Markov processes with states expressing formations survival status and locations. The blue and red teams play the roles of cost-function minimizer and maximizer, respectively. Based on the assumption of known transition matrices, the worst-case minimization objective of the blue team is formulated as a finite-time optimization, which is solved by means of a dynamic programming equation with value function evolving according to a graph of feasible paths. Once the optimization problem is solved, the resulting decision policy takes the form of a lookup table. Online implementation necessitates that formations share information through a communications network. Real-time simulations show that the cooperative path planning and weapons management policy provides, on average, an improvement in performance when compared with single-formation routing. Furthermore, this high-level decision policy integrates seamlessly with robust formation flight control and decentralized team-level fault detection schemes proposed recently by the authors.