Robust direction-of-arrival estimation for a target speaker based on multi-task U-net based direct-path dominance test

Abstract

A dedicated multi-task network for a direct-path dominance test has been proposed recently, based on which the robustness of the direction-of-arrival estimation for a target speaker tends to be significantly improved. In this Letter, the network is further refined to avoid the original two-stage processing. Moreover, the benefit and generalization of the multi-task network are confirmed by comparison with a single-task network on a significantly larger database. The efficacy of the proposed method in high noisy environments is also validated in real application environments.

Variational Bayesian Gaussian mixture model for off‐grid DOA estimation

“…The expectation of lnπ k and ln(1 − π k ) can be given as (16) where (•) denotes the digamma function.…”

“…Some studies included non-line-of-sight (NLOS) components within the cascaded channels of reconfigurable intelligent surfaces (RIS), and together with sensor noise, classified them as white noise [13,14]. Likewise, in scenarios based on acoustic signals [15,16], researchers categorized both reverberation and sensor noise as white noise. In variable inference, SBLRV and CGDP-SBL2 update variables using expectation maximization (EM), where EM, compared to mean field (MF) [17] utilized in OGSBI, lacks variance information, reducing estimation accuracy.…”