A Comparison of Axiomatic Distance-Based Collective Intelligence Methods for Wireless Sensor Network State Estimation in the Presence of Information Injection

摘要

Wireless sensor networks are a cost-effective means of data collection, especially in areas which may not have significant infrastructure. There are significant challenges associated with the reliability of measurements, in particular due to their distributed nature. As such, it is important to develop methods that can extract reliable state estimation results in the presence of errors. This work proposes and compares methods based on collective intelligence ideas, namely consensus ranking and rating models, which are founded on axiomatic distances and intuitive social choice properties. The efficacy of these methods to assess a transmitted signal’s strength with varying quantity and quality of incompleteness in the network’s readings is tested.