A Networked Decision and Information System (NDIS) architecture is proposed for the routing and munitions management (RMM) of multiple Unmanned Combat Vehicles (UCVs) evolving in an imperfectly known and adversarial environment. Increased UCV teaming agility is evidenced by the online decision policy proposed in this paper. The policy exploits the multiformations capability of UCVs to group into large formations and to divide into smaller formations on their way to high-value, tactical targets. The NDIS architecture is composed of two principal components: (i) a sensory information management network (SIM-Net), which handles data from a set of mobile sensors and then determines whether sensor redundancy should and can be used, and estimates an information state vector on the locations of the adversarial ground units and decoys; (ii) the networked UCVs calculate the worst-case minimization policy for the RMM problem based on the available information state vector. NDIS adopts a distributed one-step lookahead approach, thereby enabling time-constrained approximations of the expected cost-to-go function.
展开▼