Lower-norm criterion based background noise estimation for simple observation model

Akitoshi Itai, Yuta Hara

Abstract: This paper shows a novel estimation algorithm based on the outer product expansion with lower norms for the background noise. We have proposed a blind source separation using an outer product expansion with L1 norm minimization. The effectiveness of outer product expansions for artificial signals and an electromagnetic wave data was represented. However, the estimation performance is decreasing with an increasing of local signals. In this paper, we propose the outer product expansion with lower norms (0.1 ~ 0.9). Simulation results show that the proposed method produces the accurate background noise estimation.