Optimal distributed minimum-variance beamforming approaches for speech enhancement in wireless acoustic sensor networks

Abstract

In multiple speaker scenarios, the so-called linearly constrained minimum variance (LCMV) beamformer is a popular microphone array-based speech enhancement technique, as it allows minimizing the noise power while maintaining a set of desired responses towards the different speakers. In this paper, we address the algorithmic challenges arising in the application of the LCMV beamformer in so-called wireless acoustic sensor networks (WASNs), which are a next-generation technology for audio acquisition and processing. We review three optimal distributed LCMV-based algorithms, which compute a networkwide LCMV beamformer output at each node without centralizing the microphone signals. Optimality here refers to the fact that the algorithms theoretically generate the same beamformer outputs as in a centralized realization where a single processor would have access to all the signals. We derive and motivate the algorithms in an accessible top-down framework that reveals the underlying relations between them, as well as their differences. We explain how these differences result from their different design criterion (node-specific versus common constraints sets), as well as their different priorities with respect to communication bandwidth, computational power, and adaptivity. Furthermore, although the three algorithms were originally proposed for a fully-connected WASN, we also explain how they can be extended to the case of a partially-connected WASN, which is assumed to be pruned to a tree topology. Finally, we discuss the advantages and disadvantages of the various algorithms.

Room geometry estimation from acoustic echoes using graph-based echo labeling

“…However, this reduction comes at the cost of a decrease in the rate of adaptation of DGSC to changes in the noise field. As such, in practical applications DGSC requires a more frequent updating scheme to improve its ability to track noise correlation, increasing the total power it consumes [10].…”

“…This results in slower convergence to the optimal BF even in the case of static noise fields. Additionally, as only one node updates per audio section in DLCMV, increasing the number of nodes in the network decreases the algorithms ability to adapt to changing noise fields [10]. To demonstrate this, we consider the case of a linear array with mi = 1 ∀i ∈ V targeting a single speaker positioned perpendicular to the array.…”

Multi-task wireless acoustic sensor network for node-specific speech enhancement and DOA estimation

“…• If all nodes were MVDR nodes, i.e., if K = K MVDR , one could run the linearly constrained (LC-) DANSE algorithm [33], in which all nodes sequentially perform…”

“…In all of the three aforementioned algorithms (i.e., [25], [33], [20]), the fusion rule is updated in a similar fashion, i.e., the updating node q updates fq by replacing it with the first Mq rows of wq, i.e.,…”

Phase reference for the generalized multichannel Wiener filter

“…Recently, research on speech enhancement using socalled acoustic sensor networks consisting of spatially distributed microphones has gained significant interest [1][2][3][4][5][6][7][8][9][10][11][12]. Compared with a microphone array at a single position, spatially distributed microphones are able to acquire more information about the sound field.…”