Recursive State Estimation for Stochastic Complex Networks Under Round-Robin Communication Protocol: Handling Packet Disorders

• Published in: IEEE transactions on network science and engineering Vol. 8; no. 3; pp. 2455 - 2468
• Main Authors: Liu, Dan, Wang, Zidong, Liu, Yurong, Alsaadi, Fawaz E., Alsaadi, Fuad E.
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.07.2021; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Communication; Complex networks; Data transmission; Delays; Design of experiments; Disorders; Error analysis; mean-square boundedness; packet disorders; Parameter estimation; Protocols; Random variables; Recursive state estimation; Round-Robin protocol; Signal processing; State estimation; Stochastic processes; Time measurement; Upper bound; Upper bounds
• ISSN: 2327-4697; 2334-329X
• DOI: 10.1109/TNSE.2021.3095217

This paper investigates the recursive state estimation problemfor a class of discrete-time stochastic complex networks with packet disorders under Round-Robin (RR) communication protocols. The phenomenon of packet disorders results from the random transmission delays during the signal propagation process due to the unpredictable fluctuations of the network load, and such random delays are modeled by a set of random variables satisfying certain known probability distributions. For the sake of lessening the communication burden and abating the data collisions, the RR protocol is introduced to govern the order of the nodes for data transmission. Under the scheduling of the RR protocol, only one node is allowed to gain the access to the network at each time instant. Then, a recursive estimator is devised to guarantee an upper bound for the estimation error covariance, and then the obtained upper bound is locally minimized by adequately choosing the estimator parameters. Furthermore, the boundedness of estimation error is analyzed in the sense of mean square with the help of stochastic analysis techniques. At last, a simulation example is presented to show the applicability of the proposed estimator design scheme.