Inverse Acoustic Methodology for Continuous-Scan Phased Arrays

The paper presents a methodology for the direct estimation of the spatiospectral distribution of an acoustic source from microphone measurements that comprise fixed and continuously scanning sensors. The nonstationarity introduced by the sensor motion is quantified by means of the Wigner-Ville spectrum. Its strongest effect is on the correlations of the sensor signals. Suppression of the nonstationarity in the signal processing involves division of the signals into blocks and application of a frequency-dependent window within each block. The direct estimation approach entails the inversion of an integral that connects the source distribution to the measured coherence of the acoustic field. A Bayesian estimation approach is developed that allows for efficient inversion of the integral and performs similarly to the much costlier conjugate-gradient method. The methodology is applied to acoustic fields emitted by impinging jets approximating a point source and an overexpanded supersonic jet. The measurement setup comprises one continuously scanning microphone and a number of fixed microphones, which are all arranged on a linear array. Comparisons are made between array configurations with fixed microphones only and with the scanning microphone: all having the same sensor count. The noise source maps with the scanning microphone have improved spatial fidelity and suppressed side lobes. The ability of the continuous-scan paradigm to provide high-definition noise source maps with a low sensor count is demonstrated.