Dynamic Replanning for Multi-UAV Persistent Surveillance
Nikhil Kumar Nigam · AIAA Guidance, Navigation, and Control (GNC) Conference · 2013
Swarms of autonomous vehicles are increasingly being considered for several applications, including weather monitoring, geographical surveys and extra-terrestrial exploration. The task of persistent surveillance is particularly relevant in situations where the target area needs to be continuously surveyed, minimizing the time between visitations to the same region. This distinction from one-time coverage does not allow a straightforward application of most exploration techniques to the problem, especially in presence of vehicle dynamic constraints. Furthermore, the stochastic nature of the environment coupled with unanticipated failures makes the mission management and coordination problem significantly challenging. In this research, we investigate techniques for high-level control, that are scalable, reliable, efficient, and robust to problem dynamics, and allow near real-time replanning in unscripted environments. We present a decentralized hybrid discrete-continuous hierarchical approach for combined control and mission planning, building upon our previous work on semi-heuristic multiple Unmanned Air Vehicle (UAV) control policies and enhancing it with a Probability Collectives (PC) based approach for dynamic mission replanning. PC combines bounded rationality game theory with statistical physics, using information theory. As such, it provides an efficient mechanism for learning in games and provides a single framework for treatment of both continuous and discrete variables/decisions. In this paper we focus on demonstrating the ability of the algorithm approach to plan and replan efficiently replan while maintaining performance.