||Eglin Air Force Base, FL 325426810
The utilization of autonomous agents can support mission success in dynamic, uncertain, and contested environments by augmenting human operator capabilities. Specifically, cooperative autonomous systems can provide enhanced situational awareness, which can inform decision support for multi-target search, track, and engagement scenarios. To leverage the full capabilities of autonomous agents in a dynamic and uncertain battlefield, a common framework to rapidly update situational awareness and task assignments between all agents, both human and autonomous, is required.
This need for enhanced situational awareness across a team is also challenged by dynamic and uncertain information and communication topologies in decentralized command and control architectures. It has been shown that communicating the uncertainty of an autonomous team’s estimate of the state of the environment enables a human operator to make more informed and risk-aware decisions. As such, the bi-directional communication of the uncertainty of information passed between agents and the ability to update one's own world model is vital to a shared understanding. The goal of this research is to address the aforementioned challenges to enable the synergistic teaming of heterogeneous agents to search for, track, and engage multiple targets in complex, contested, and dynamic environments.
Addressing such challenges requires a multi-disciplinary approach including but not limited to (1) probabilistic methods for target state prediction and multi-agent control, (2) graph theoretic representations of communication and command topologies, (3) optimization techniques to aid in mission planning given the predicted target set and the individual and joined state information within an autonomous team, and (4) experimental demonstrations of proposed techniques.