Adaptive Identification of Acoustic Multichannel Systems Using Sparse Representations

The book "Adaptive Identification of Acoustic Multichannel Systems Using Sparse Representations" addresses the extension of adaptive filter theory in the context of massive multichannel systems. It takes into account a priori knowledge about the underlying system or signal. The focus is on leveraging sparsity in acoustic multichannel systems to solve the problem of non-uniqueness with an efficient algorithm for adaptive filtering that does not require modification of the speaker signals. The work provides a detailed derivation of general sparse representations of acoustic MIMO systems in signal- or system-dependent transformation domains. Efficient adaptive filter algorithms are presented in these domains, highlighting the relationship between signal- and system-based sparse representations. Additionally, a novel approach to spatial preprocessing of speaker signals in a full-duplex communication system is introduced to prevent echo capture by the microphone array and to support the AEC system. The book concludes with the presentation of a multichannel system for acoustic echo cancellation, which serves as a post-processing stage to remove residual echoes.