Distributed Frequency Estimation Over Sensor Network

Abstract

Frequency estimation is a common problem in a variety of applications. In recent years, adaptive notch filtering methods have been widely adopted for solving frequency estimation problems. For frequency estimation performed by a single node, the sampling time may not be long enough, the sampling rate may not be high enough, and the noise effect may be serious. In such situations, most existing algorithms for frequency estimation can not produce accurate enough results. Here, we propose using wireless sensor networks for sampling data distributedly, and using distributed notch filtering method for dealing with this problem. In particular, we propose two distributed algorithms over sensor network, a least mean square-based distributed notch filtering (dNF) algorithm and a total least square-based dNF algorithm. The communication cost of the new proposed algorithms is low, as each node exchanges only with its neighbors the estimates other than the original data. The proposed algorithms are applied to both synthetic and real examples. Simulation results demonstrate the effectiveness of the new proposed algorithms.