Decentralized adaptive tracking with one-step-guess estimator

Decentralized adaptive tracking problem for a simple yet typical discrete-time multi-agent system is investigated, where each agent is affected by its neighborhood agents with unknown interaction strength. In this paper, for the purpose of decentralized adaptive tracking, all agents are supposed to be smart enough with limited intelligence such that they can adopt the most simple one-step-guess (OSG) algorithm to estimate the unknown interaction strength and then design local adaptive controllers with the certainty-equivalence principle. The whole closed-loop system is shown to be rather complex and it has been shown that, under some mild conditions, the closed-loop stability is established in the case of bounded noise, while asymptotic tracking can be established in the case of diminishing noise. Extensive simulations verified correctness of theoretical results.