Low Complexity Signal Adaptive Sound Zone Control using Subspace Tracking

Sound zone control methods usually do not consider the spectral characteristics of the input signal, or only consider long-term averaged spectral characteristics. This makes the control effect optimal only in an average sense or only for white signals. In this work, we show that taking into account the instantaneous spectral characteristics in the acoustic contrast control (ACC) method (inst-ACC) can improve the acoustic contrast compared to existing approaches, but requires high computational complexity. To reduce complexity, we formulate a recursive version of the inst-ACC (R-inst-ACC), where the required eigenvector is obtained using the method of projection approximation subspace tracking with deflation (PASTd). Using the proposed method, the complexity is reduced from $\mathcal{O}(L^3J^3)$ to $\mathcal{O}(M_D L^2J^2) $,
with $LJ$ being the length of the control filter vector and $M_D \ll LJ$ being the number of control points in the dark zone. Simulation results show the effectiveness of the proposed R-inst-ACC algorithm.