This paper considers the problem of path planning for a team of unmanned aerial vehicles performing surveillance near a friendly base. The unmanned aerial vehicles do not possess sensors with automated target recognition capability and, thus, rely on communicating with unattended ground sensors placed on roads to detect and image potential intruders. The problem is motivated by persistent intelligence, surveillance, reconnaissance and base defense missions. The problem is formulated and shown to be intractable. A heuristic algorithm to coordinate the unmanned aerial vehicles during surveillance and pursuit is presented. Revisit deadlines are used to schedule the vehicles' paths nominally. The algorithm uses detections from the sensors to predict intruders' locations and selects the vehicles' paths by minimizing a linear combination of missed deadlines and the probability of not intercepting intruders. An analysis of the algorithm's completeness and complexity is then provided. The effectiveness of the heuristic is illustrated through simulations in a variety of scenarios.

In this paper, we model operator states using hidden Markov models applied to human supervisory control behaviors. More specifically, we model the behavior of an operator of multiple heterogeneous unmanned vehicle systems. The hidden Markov model framework allows the inference of higher operator states from observable operator interaction with a computer interface. For example, a sequence of operator actions can be used to compute a probability distribution of possible operator states. Such models are capable of detecting deviations from expected operator behavior as learned by the model. The difficulty with parametric inference models such as hidden Markov models is that a large number of parameters must either be specified by hand or learned from example data. We compare the behavioral models obtained with two different supervised learning techniques and an unsupervised hidden Markov model training technique. The results suggest that the best models of human supervisory control behavior are obtained through unsupervised learning. We conclude by presenting further extensions to this work.

This paper considers a path planning problem with two marsupial vehicles (one carrier vehicle and one passenger vehicle that is deployed by the carrier vehicle) exploring a planar area. This work is motivated by multiagent intelligence, surveillance, and reconnaissance missions in contested environments. The vehicles are heterogeneous, e.g., the carrier vehicle is faster than the passenger vehicle or the passenger vehicle possesses better sensors than the carrier vehicle. The vehicles are to gather a finite amount of information about an object of interest using their sensors while minimizing the likelihood of their detection by an opponent. Necessary conditions for optimal solutions are given and straight line travel for both vehicles is shown to be optimal. The results are illustrated in several simulation examples.

No abstract