Optimal sensor scheduling for hybrid estimation

Abstract

A sensor scheduling problem was considered for a class of hybrid systems named as the stochastic linear hybrid system (SLHS). An algorithm was proposed to select one (or a group of) sensor at each time from a set of sensors. Then, a hybrid estimation algorithm was designed to compute the estimates of the continuous and discrete states of the SLHS based on the observations from the selected sensors. As the sensor scheduling algorithm is designed such that the Bayesian decision risk is minimized, the true discrete state can be better identified. Moreover, the continuous state estimation performance of the proposed algorithm is better than that of hybrid estimation algorithms using only predetermined sensors. Finally, the algorithms are validated through an illustrative target tracking example.

Keywords

sensor scheduling / hybrid systems / Bayesian decision risk / target tracking

References

[1] KrishnamurthyV, DjoninD V. Optimal threshold policies for multivariate POMDPs in radar resource management [J]. IEEE transactions on Signal Processing, 2009, 57: 3954-3969

[2] DoshiB, HenrickB, BenmohamedL, ChimentoP, WangIRossileH. Sensor network design for underwater surveillance [C]. Military Communications Conference, 2006Washington D.C, USAAcademic Press Corporation1-7

[3] StevensT J, MorrellD. Minimization of sensor usage for target tracking in a network of irregularly spaced sensors [C]. Robert B.IEEE. Workshop on Statistical Signal Processing, 2003St Louis, USAAcademic Press Corporation539-542

[4] PeschonJ, MeierLIII, DresslerR. Optimal control of measurement subsystems [J]. IEEE Transactions on Automatic Control, 1967, 12: 528-536

[5] HerringK, MelsaJ. Optimum measurements for estimation [J]. IEEE Transactions on Automatic Control, 1974, 19(3): 264-266

[6] BarasJ S, BensoussanA. Optimal sensor scheduling in nonlinear filtering of diffusion processes [J]. SIAM J Control Optim, 1989, 27(4): 786-813

[7] AthansM. On the determination of optimal costly measurement strategies [J]. Automatica, 1972, 18: 397-412

[8] ShinJ, ZhaoF, ReichJ. Information-driven dynamic sensor collaboration for tracking applications [J]. IEEE Signal Processing Magazine, 2002, 19(2): 61-72

[9] SinopliB, ShiL, EpsteinM, MurrayR MJustinR M. Effective sensor scheduling schemes in a sensor network by employing feedback in the communication loop [C]. IEEE International Conferenceon Control Application, 2007SingaporeSouth Ocean Publishing House1006-1011

[10] HassibiB, GuptaV, ChungT H, MurrayR M. On a stochastic sensor selection algorithm with applications in sensor scheduling and sensor coverage [J]. Automatica, 2006, 42: 251-260

[11] GuptaV, HovareshtiP, BarasJ. On sensor scheduling using smart sensors [C]. Jim Spall. In Proc. of the IEEE Conference on Decision and Control (CDC’ 07), 2007Louisiana, USASpringer-Verlag494-499

[12] FranceschettiM, PoollaK, SchenatoL, SinopoliB, SastryS. Foundations of control and estimation over lossy networks[C]. Panos, John Baillieul. Proceedings of the IEEE, Special issue on Networked Control Systems, 2007Grand Wailea, USASpringer-Verlag163-181

[13] HeY, ChongE K P. Sensor scheduling for target tracking: A montecarlo sampling approach [J]. Digital Signal Processing, 2006, 2: 533-545

[14] OshmanY. Optimal sensor selection strategy for discrete-time state estimator [J]. IEEE Transactions on Aerospace and Electronic Systems, 1994, 30(2): 307-314

[15] ChongE K P, LiY, KrakowL W, GroomK N. Approximate stochastic dynamic programming for sensor scheduling to track multiple targets [J]. Digital Signal Processing, 2009, 19: 978-989

[16] HallD L, LlinasJFranchetiL B. An introduction to multi-sensor data fusion [C]. Proceedings of the IEEE on Decision and Control (CDC’ 08), 1997USAJohn Wiley & Sons6-23

[17] AbateA, HuJ, VitusM P, ZhangW, TomlinCPollaK. On efficient sensor scheduling for linear dynamical systems [C]. In Proc. American Control Conference, 2010Baltimore, USANational Academic Press4833-4838

[18] BemporadA, BernardiniD, MunozD, PenaD, FrazzoliE. Simultaneous optimal control and discrete stochastic sensor selection [C]. Cochron. In Hybrid System Computation and Control (HSCC) Conference, 2009USASpringer-Verlag61-75

[19] Bar-ShalomY, LiX R, KirubarajanTEstimation with applications to tracking and navigation [M], 2001New YorkJohn Wiley & Sons65-72

[20] HartP, DudaR, StorkDPattern classification [M], 2000New YorkWiley Interscience172-176

[21] RajamanoharanS, BlackmoreL, WilliamsB C. Active estimation for jump Markov systems [J]. IEEE Transactions on Automatic Control, 2008, 53(10): 2223-2236

[22] SeahC E, HwangI. Stochastic linear hybrid systems: Modeling, estimation, and application in air traffic control [J]. IEEE Transactions on Control Systems Technology, 2009, 17(3): 563-575

[23] LiX R, Bar-ShalomY. Design of an interacting multiple model algorithm for air traffic control tracking [J]. IEEE Transactions on Control Systems Technology, 1993, 1(3): 186-194

[24] Bar-ShalomY, FortmannT FTracking and Data Association [M], 1988New YorkAcademic Press125-128