Distributed Spatial Filtering by a Two-Hop Consensus-Type Algorithm

In this letter, we discuss distributed spatial filtering (DSF) on networked systems to obtain signal values with a desired spatial frequency characteristic from those assigned to nodes by a distributed algorithm. We present a two-hop consensus-type algorithm for DSF based on an existing one-hop algorithm. We prove that the range of the filter characteristics the presented algorithm can achieve is broader than that for the existing algorithm by deriving a necessary and sufficient condition for achieving DSF. Simulation results show that our filtering algorithm and a new filter characteristic it provides are effective in distributed anomaly detection by sensor networks.