Decision making based on localized auctions in wireless sensor networks

Wireless sensor networks are often only used to transmit monitored data to a central processing unit. However, they also offer the possibility to establish local intelligence and allow localized problem solutions. Existing monitoring systems have the following main drawbacks: centralized architecture that requires expensive infrastructure to run it, centralized decision making that makes reactions on events too slow and has high communication cost. In this paper we explore application of localized auction algorithms in decision making process. We show advantages of localized decision making over centralized in terms of communication cost minimization. Instead of using usual decision-level methods (e.g. weighted decision, classical inference etc.) we propose using localized auctions to get and combine relevant sensor information to make the decision. Simulation results are shown in the paper.