Wideband Source Localization Using One-Bit Quantized Arrays

We consider the problem of source localization with signals that are quantized to a single bit. While this problem has been previously addressed for narrowband signals quantized in complex baseband, we consider wideband signals quantized directly in the time domain. We adapt the generalized cross correlation estimator to work with quantized signals and analyze its theoretical performance as a function of the signal-to-noise ratio (SNR). We find that quantization imposes a small penalty at low SNR but scales poorly at high SNR. Furthermore, we analyze the effects of sampling on estimation performance and show that oversampling is necessary at high SNR. Finally, we apply the generalized cross correlation weighting function to steered response power direction-of-arrival estimation with a larger array. We demonstrate the performance of the system for speech localization.