1092 IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS—PART A: SYSTEMS AND HUMANS, VOL. 41, NO. 6, NOVEMBER 2011 Hidden Markov Model and Auction-Based Formulations of Sensor Coordination Mechanisms in Dynamic Task Environments Woosun An, Chulwoo Park, Member, IEEE, Xu Han, Krishna R. Pattipati, Fellow, IEEE, David L. Kleinman, Fellow, IEEE, and William G. Kemple Abstract—In this paper, multistage auction-based intelligence, surveillance, and reconnaissance (ISR) sensor coordination mech- anisms are investigated in the context of dynamic and uncertain mission environments such as those faced by expeditionary strike groups. Each attribute of the mission task is modeled using a hidden Markov model (HMM) with controllable emission matri- ces, corresponding to each ISR asset package (subset of sensors). For each HMM-asset package pair, we evaluate a matrix of in- formation gains (uncertainty reduction measures). The elements of this matrix depend on the asset coordination structure and the concomitant delays accrued. We consider three coordination structures (distributed ISR coordination, ISR officer serving as a coordinator, and ISR officer serving as a commander) here. We evaluate these structures on a hypothetical mission scenario that requires the monitoring of ISR activities in multiple geographic regions. The three structures are evaluated by comparing the task state estimation error cost, as well as travel, waiting, and assignment delays. The results of the analysis were used as a guide in the design of a mission scenario and asset composition for a team-in-the-loop experimentation. Our solution has the potential to be a mixed initiative decision support tool to an ISR coordi- nator/commander, where the human provides possible ISR asset package-task pairings and the tool evaluates the efficacy of the assignment in terms of task accuracy and delays. We also apply our approach to a hypothetical disaster management scenario involving chemical contamination and discuss the computational complexity of our approach. Index Terms—Auction algorithm, coordination delays, hidden Markov model (HMM), information gain (IG) heuristic sensor scheduling, partition algorithm, sensor assignment. Manuscript received August 12, 2009; revised March 22, 2010; accepted September 4, 2010. Date of publication March 24, 2011; date of current version October 19, 2011. This work was supported by the Office of Naval Research under Grant N00014-09-1-0062. This paper was recommended by Associate Editor L. C. Jain. W. An, C. Park, X. Han, and K. R. Pattipati are with the Electrical Engineering Department, University of Connecticut, Storrs, CT 06029 USA (e-mail: woa05001@engr.uconn.edu; chp06004@engr.uconn.edu; krishna@ engr.uconn.edu). D. L. Kleinman and W. G. Kemple are with the Department of Information Sciences, Naval Postgraduate School, Monterey, CA 93943 USA (e-mail: dlkleinm@nps.edu; kemple@nps.edu). Color versions of one or more of the figures in this paper are available online at http://ieeexplore.ieee.org. Digital Object Identifier 10.1109/TSMCA.2011.2114342 I. I NTRODUCTION C OMPLEX surveillance applications, such as the use of un- manned aerial vehicles (UAVs) for monitoring activities in remote or hostile environments, require one to tradeoff the sensor performance (e.g., detection, identification, and tracking accuracies) and the sensor usage cost (e.g., power and band- width consumption, distance traveled, risk of exposure, and deployment requirements). The objective of dynamic sensor scheduling is to judiciously allocate sensing resources to ex- ploit the individual sensors’ capabilities while minimizing their usage cost. UAVs are preferred assets for monitoring nearly all of the intelligence, surveillance, and reconnaissance (ISR) activities; however, they cannot be deployed in large numbers due to their limited availability. Thus, astute allocation of scarce resources is a major issue in ISR coordination. In addition, sensors have different measurement capabilities with varying accuracies. Thus, the sensor scheduling algorithm needs to coordinate sensors and to tradeoff accuracy versus timelines of measurement as well as risk of exposure. In this paper, we develop analytic models of an expeditionary strike group (ESG) with different ISR coordination structures tasked with executing a surveillance mission. An ESG provides a flexible Navy–Marine force, which is capable of tailoring itself to a wide variety of missions. An important ESG mission involves surveillance and prosecution of asymmetric threats. The increased geographical range and the unpredictable nature of their threat behavior require effective allocation and appro- priate scheduling of sensors to achieve the mission objectives. Effectively performing the ISR activities is a key step to gain situational awareness, which, in turn, enables the allocation of resources for the interdiction of asymmetric threats. We model the asymmetric threats using hidden Markov models (HMMs) because these activities are concealed and their true state can only be inferred through the uncertain observations obtained using various ISR sensors. A pattern of these observations and its dynamic evolution over time provides the information base for inferring a potential realization of an asymmetric threat. Thus, each state of an HMM is characterized by a set of attributes, and a sensor package consisting of a subset of sensors is needed to accurately estimate these attributes and, consequently, to infer the task (HMM) state. This is the type of problem considered by Hutchins et al. [1], where they have examined how an ESG with alternative 1083-4427/$26.00 © 2011 IEEE