Simultaneous Perturbation Stochastic Approximation-Based Consensus for Tracking Under Unknown-But-Bounded Disturbances

被引:17
作者
Granichin, Oleg [1 ]
Erofeeva, Victoria [1 ]
Ivanskiy, Yury [1 ]
Jiang, Yuming [2 ]
机构
[1] St Petersburg State Univ, Sci & Educ Ctr Math Robot & Artificial Intelligen, St Petersburg 198504, Russia
[2] Norwegian Univ Sci & Technol NTNU, Dept Informat Secur & Commun Technol, NO-7491 Trondheim, Norway
基金
俄罗斯科学基金会;
关键词
Sensors; Approximation algorithms; Optimization; Noise measurement; Perturbation methods; Network topology; Upper bound; Arbitrary noise; consensus algorithm; distributed tracking; multiagent networks; randomized algorithm; simultaneous perturbation stochastic approximation (SPSA); stochastic stability; tracking performance; unknown-but-bounded disturbances;
D O I
10.1109/TAC.2020.3024169
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
We consider a setup where a distributed set of sensors working cooperatively can estimate an unknown signal of interest, whereas any individual sensor cannot fulfill the task due to lack of necessary information diversity. This article deals with these kinds of estimation and tracking problems and focuses on a class of simultaneous perturbation stochastic approximation (SPSA)-based consensus algorithms for the cases when the corrupted observations of sensors are transmitted between sensors with communication noise and the communication protocol has to satisfy a prespecified cost constraints on the network topology. Sufficient conditions are introduced to guarantee the stability of estimates obtained in this way, without resorting to commonly used but stringent conventional statistical assumptions about the observation noise, such as randomness, independence, and zero mean. We derive an upper bound of the mean square error of the estimates in the problem of unknown time-varying parameters tracking under unknown-but-bounded observation errors and noisy communication channels. The result is illustrated by a practical application to the multisensor multitarget tracking problem.
引用
收藏
页码:3710 / 3717
页数:8
相关论文
共 26 条
[1]   Efficient sensor management policies for distributed target tracking in multihop sensor networks [J].
Aeron, Shuchin ;
Saligrama, Venkatesh ;
Castanon, David A. .
IEEE TRANSACTIONS ON SIGNAL PROCESSING, 2008, 56 (06) :2562-2574
[2]   Approximate Consensus in Stochastic Networks With Application to Load Balancing [J].
Amelina, Natalia ;
Fradkov, Alexander ;
Jiang, Yuming ;
Vergados, Dimitrios J. .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2015, 61 (04) :1739-1752
[3]   ASYMPTOTIC AGREEMENT IN DISTRIBUTED ESTIMATION [J].
BORKAR, V ;
VARAIYA, PP .
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1982, 27 (03) :650-655
[4]  
Borkar Vivek, 2009, Stochastic Approximation: A Dynamical Systems Viewpoint, V48
[5]   Distributed optimization and statistical learning via the alternating direction method of multipliers [J].
Boyd S. ;
Parikh N. ;
Chu E. ;
Peleato B. ;
Eckstein J. .
Foundations and Trends in Machine Learning, 2010, 3 (01) :1-122
[6]   Randomized gossip algorithms [J].
Boyd, Stephen ;
Ghosh, Arpita ;
Prabhakar, Balaji ;
Shah, Devavrat .
IEEE TRANSACTIONS ON INFORMATION THEORY, 2006, 52 (06) :2508-2530
[7]   A Kiefer-Wolfowitz algorithm with randomized differences [J].
Chen, HF ;
Duncan, TE ;
Pasik-Duncan, B .
IEEE TRANSACTIONS ON AUTOMATIC CONTROL, 1999, 44 (03) :442-453
[8]   REACHING A CONSENSUS [J].
DEGROOT, MH .
JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 1974, 69 (345) :118-121
[9]   NEXT: In-Network Nonconvex Optimization [J].
Di Lorenzo, Paolo ;
Scutari, Gesualdo .
IEEE TRANSACTIONS ON SIGNAL AND INFORMATION PROCESSING OVER NETWORKS, 2016, 2 (02) :120-136
[10]  
Erofeeva V, 2019, IEEE DECIS CONTR P, P6050, DOI 10.1109/CDC40024.2019.9030129