Decentralized adaptive control using an affine plus self-organizing fuzzy neural network for Multi-Agent System consensus problem
Masanao Obayashi, Yasuhiro Otomi, Takashi Kuremoto, Kunikazu Kobayashi, Shingo Mabu · 2013
In this paper, we propose an effective method to configure a dynamical structure of each agent constituting Multi-Agent System (MAS) on a decentralized adaptive control. It is important that each agent does decision-making while configuring its own desirable dynamical characteristics and adapting to environmental changes. In conventional researches, the dynamics of each agent is modeled by neural network (NN) with static structure. Therefore, it is difficult for the agent to behave appropriately at time-varying conditions due to the static structure of NN. Thus, we propose a new decentralized adaptive control system (DACS) using an affine plus self-organizing fuzzy neural network (ASOFNN) for MAS, considering the consensus problem. Additionally, we give the proof of stability analysis of the proposed method theoretically, and the effectiveness of the proposed method is verified by the computational simulations.