×

Heavy-tails in Kalman filtering with packet losses. (English) Zbl 1425.93287

Summary: In this paper, we study the existence of a steady-state distribution and its tail behaviour for the estimation error arising from Kalman filtering for unstable linear dynamical systems. Although a large body of literature has studied the problem of Kalman filtering with packet losses in terms of analysis of the second moment, no study has addressed the actual distribution of the estimation error. First we show that if the system is strictly unstable and packet loss probability is strictly less than unity, then the steady-state distribution (if it exists) must be heavy tail, i.e. its absolute moments beyond a certain order do not exist. Then, by drawing results from renewal theory, we further provide sufficient conditions for the existence of such stationary distribution. Moreover, we show that under additional technical assumptions and in the scalar scenario, the steady-state distribution of the Kalman prediction error has an asymptotic power-law tail, i.e. \(P[|e|>s]\sim s^{-\alpha}\), as \(s\rightarrow\infty\), where \(\alpha\) can be explicitly computed. We further explore how to optimally select the sampling period assuming an exponential decay of packet loss probability with respect to the sampling period. In order to minimize the expected value of the second moment or the confidence bounds, we illustrate that in general a larger sampling period will need to be chosen in the latter case as a result of the heavy tail behaviour.

MSC:

93E11 Filtering in stochastic control theory
93C15 Control/observation systems governed by ordinary differential equations
68M10 Network design and communication in computer systems
94A12 Signal theory (characterization, reconstruction, filtering, etc.)
60H15 Stochastic partial differential equations (aspects of stochastic analysis)
PDFBibTeX XMLCite
Full Text: DOI

References:

[1] Antonello, R.; Schenato, L., Laboratory activity 4: Longitudinal statespace control of the balancing robot, Technical report (2017), University of Padova, Department of Information Engineering, http://automatica.dei.unipd.it/tl_files/utenti/lucaschenato/SEGWAY_GUIDE.pdf
[2] Brandt, A., The stochastic equation \(y_{n + 1} = a_n y_n + b_n\) with stationary coefficients, Adv. Appl. Prob., 18, 211-220 (1986) · Zbl 0588.60056
[3] Buraczewski, D.; Damek, E.; Mikosch, T., Stochastic Models with Power-Law Tails: The Equation \(X = A X + B (2016)\), Springer International Publishing · Zbl 1357.60004
[4] Censi, A., Kalman filtering with intermittent observations: Convergence for semi-Markov chains and an intrinsic performance measure, IEEE Trans. Autom. Control, 56, 2, 376-381 (2010) · Zbl 1368.93707
[5] Dey, S.; Schenato, L., Heavy-tails in Kalman filtering with packet losses: confidence bounds vs. second moment stability, Proceedings of the European Control Conference (2008)
[6] Goldie, C., Implicit renewal theory and tails of solutions of random equations, Ann. Appl. Probab., 1, 1, 126-166 (1991) · Zbl 0724.60076
[7] Goldie, C.; Grubel, R., Perpetuities with thin tails, Adv. Appl. Prob., 28, 463-480 (1996) · Zbl 0862.60046
[8] Goldsheid, I.; Margulis, G., Lyapunov indices of a product of random matrices, Russian Math. Surv., 44, 11-71 (1989) · Zbl 0705.60012
[9] Grincevicius, A., One limit distribution for a random walk on the line, Lith. Math. J., 15, 580-589 (1975) · Zbl 0373.60009
[10] Hespanha, J.; Naghshtabrizi, P.; Xu, Y., A survey of recent results in networked control systems, Proc. IEEE, 95, 1, 138-162 (2007)
[11] Hitczenko, P.; Wesolowski, J., Perpetuities with thin tails, revised, Ann. Appl. Probab., 19, 2080-2101 (2009) · Zbl 1185.60074
[12] Huang, M.; Dey, S., Stability of Kalman filtering with Markovian packet losses, Automatica, 43, 598-607 (2007) · Zbl 1261.93083
[13] Kar, S.; Sinopoli, B.; Moura, J., Kalman filtering with intermittent observations: weak convergence to a stationary distribution, IEEE Trans. Autom. Control, 57, 405-420 (2012) · Zbl 1369.93630
[14] Kesten, H., Random difference equations and renewal theory for products of random matrices, Acta Math., 131, 207-248 (1973) · Zbl 0291.60029
[15] Loeve, M., Probability Theory (1960), Van Nostrand: Van Nostrand New York · Zbl 0108.14202
[16] Mannion, D., Products of 2 × 2 random matrices, Ann. Appl. Probab., 3, 4, 1189-1218 (1993) · Zbl 0784.60019
[17] Mo, Y.; Sinopoli, B., A characterization of the critical value for Kalman filtering with intermittent observations, Proceedings of the 47th IEEE Conference on Decision and Control, 2692-2697 (2008), IEEE
[18] Mo, Y.; Sinopoli, B., Kalman filtering with intermittent observations: Tail distribution and critical value, IEEE Trans. Auto. Control, 57, 3, 677-689 (2012) · Zbl 1369.93643
[19] Park, P.; Araujo, J.; Johansson, K., Wireless networked control system co-design, Proceedings of the International Conference on Networking, Sensing and Control, Delft, the Netherlands, 486-491 (2011)
[20] Park, S. Y.; Sahai, A., Intermittent Kalman filtering: Eigenvalue cycles and nonuniform sampling, Proceedings of the American Control Conference (2011)
[21] Plarre, K.; Bullo, F., On kalman filtering for detectable systems with intermittent observations, IEEE Trans. Autom. Control, 54, 2, 386-390 (2009) · Zbl 1367.93656
[22] Proakis, J.; Salehi, M., Digital Communications (2007), McGraw-Hill Education
[23] Protasov, V.; Jungers, R., Lower and upper bounds for the largest Lyapunov exponent of matrices, Linear Algebra Appl., 438, 4448-4468 (2013) · Zbl 1281.65154
[24] Rohr, E. R.; Marelli, D.; Fu, M., Kalman filtering with intermittent observations: On the boundedness of the expected error covariance, IEEE Trans. Autom. Control, 59, 2734-2738 (2010) · Zbl 1360.93710
[25] Sahai, A.; Mitter, S., The necessity and sufficiency of anytime capacity for stabilization of a linear system over a noisy communication. part I: scalar systems, IEEE Trans. Inf. Theory, 52, 8, 3369-3395 (2006) · Zbl 1309.94066
[26] de Saporta, B., Tail of the stationary solution of the stochastic equation \(y_{n + 1} = a_n y_n + b_n\) with Markovian coefficients, Stoch. Process. Appl., 115, 1954-1978 (2005) · Zbl 1096.60025
[27] Schenato, L., Kalman filtering for networked control systems with random delay and packet loss, IEEE Trans. Autom. Control, 53, 1311-1317 (2008) · Zbl 1367.93633
[28] Shi, L.; Epstein, M.; Murray, R., Kalman filtering over a packet-dropping network: a probabilistic perspective, IEEE Trans. Autom. Control, 55, 3, 594-604 (2010) · Zbl 1368.93728
[29] Shi, L.; Epstein, M.; Tiwari, A.; Murray, R., Estimation with information loss: asymptotic analysis and error bounds, Proceedings of the 44th IEEE CDC, Seville, Spain, 1215-1221 (2005)
[30] Sinopoli, B.; Schenato, L.; Franceschetti, M.; Poolla, K.; Jordan, M.; Sastry, S., Kalman filtering with intermittent observations, IEEE Trans. Autom. Control, 49, 9, 1453-1464 (2004) · Zbl 1365.93512
[31] Wu, J.; Shi, L.; Xie, L.; Johansson, K., An improved stability condition for Kalman filtering with bounded Markovian packet losses, Automatica, 62, 32-38 (2015) · Zbl 1329.93147
[32] You, K.; Fu, M.; Xie, L., Stability of Kalman filtering with Markovian packet losses, Automatica, 47, 12, 2647-2657 (2011) · Zbl 1235.93246
[33] You, K.; Xie, L., Survey of recent progress in networked control systems, Acta Sinica, 39, 2, 101-117 (2013)
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.