Noise enhanced distributed Bayesian estimation

Alireza Sani; Azadeh Vosoughi

ICASSP 2017accepted0 citations

Noise enhanced distributed Bayesian estimation

Abstract

In this paper we consider distributed estimation of an unknown Gaussian random variable with known mean and variance, where each sensor observation is affected by both multiplicative and additive Gaussian observation noises. We derive the corresponding Cramer Rao Lower Bound (CRLB) for both quantized and full precision observations. In sequel we provide some closed-form approximations for both CRLB expressions which provide us with better understanding of behavior of CRLBs. Afterwards through analytic and simulation results we report some scenarios that multiplicative observation noise can play an enhancive role in terms of estimation accuracy. We call this phenomena enhancement mode of multiplicative noise.

BibTeX

@inproceedings{icassp2017_noiseenhanceddis,
  title = {Noise enhanced distributed Bayesian estimation},
  author = {Alireza Sani and Azadeh Vosoughi},
  booktitle = {ICASSP 2017},
  year = {2017}
}