A swap-based sparse array design method in personal sound zone

Abstract

Sparse array design (SAD) optimizes the geometrical configuration of the loudspeaker array. It provides possibilities for further improving the performance of personal sound zone (PSZ). In this paper, the SAD problem in PSZ is formulated as a sparse multi-objective optimization problem, which takes multiple performance metrics into consideration during the sparsity promotion procedure. A swap-based group search algorithm is designed and proposed for solving the formulated non-convex and combinatorial problem to near-optimality. An approximate form is also proposed for reducing computational burden when there exist multiple target sound fields in bright zone. Simulation and experimental results suggest that the proposed method is more suitable in PSZ for controlling over trade-offs and achieving the same good or even better performances, and the proposed approximate form is effective.