APPLICATION OF ARTIFICIAL INTELLIGENCE METHODS FOR ADAPTING PERSONAL SOUND ZONE FORMATION SYSTEMS TO ROOM ACOUSTICS

Abstract

An analytical review of the current state of research in the field of Personal Sound Zones (PSZ) formation is presented, with an emphasis on the use of Artificial Intelligence (AI) methods for adapting such systems to complex room acoustics. A vector–matrix model of the system is introduced, and the objective functions of Acoustic Contrast Control (JACC) and Pressure Matching (JPM) algorithms are formulated with regularization parameters λ and μ. A formal analogy between the PSZ problem and beamforming in Phased Array Antennas (PAA) is established. The reasons why classical ACC and PM algorithms degrade under reverberant conditions are analyzed, and AI-based approaches aimed at solving this problem are systematized, including neural-network-based ATF identification, deep learning for nonlinear distortion compensation, and low-latency online adaptation algorithms. The review serves as a theoretical basis for dissertation research aimed at developing an AI-based adaptive PSZ control system. The conducted analysis shows that PSZ formation algorithms can be represented within a unified vector–matrix formulation, p = Hw, where the objective functions JACC and JPM formally correspond to the MVDR and LCMV beamforming criteria used in phased antenna arrays. This mathematical framework defines the point of application for artificial intelligence methods: AI does not replace classical algorithms but addresses their key vulnerability, namely their dependence on the accurate matrix H, which continuously changes under real acoustic conditions. The practical significance of the review lies in the fact that it forms the theoretical and methodological basis for dissertation research whose central task is the development of an AI-based system for adaptive PSZ control.

Keywords: localized sound zones, audio enclaves, Personal Sound Zones, loudspeaker arrays, parametric array, acoustic contrast, pressure matching, speech intelligibility, ultrasound, ultrasonic transducers, acoustic metasurfaces.