Sparse array DOA estimation in the presence of unknown non-uniform noise

This paper investigates three algorithms for estimating the direction of arrival (DOA) of incoming wireless signals. The algorithms studied are Minimum Variance Distortionless Response (MVDR), MUltiple SIgnal Classification (MUSIC) and Maximum Likelihood (ML). The goal was to study these algorithms under the following conditions: a sparse sensor array, unknown nonuniform noise and the presence of noise and signal correlation. The simulations show that while noise correlation has a limited effect on DOA estimation error, signal correlation has a strong effect, especially when ML is used. The simulations also show that MUSIC is the most effective algorithm when signal correlation exists. Tests on real signal data were also performed.^{1 2}