Abstract
Broadband source waveforms are constructed that allow for the minimum mean square error (MMSE) estimation of an observed acoustic response function with minimal computational effort at the receiver. The waveforms offers near perfect auto and cross correlation properties over the limited propagation delay spread of the acoustic response allowing for least squares estimation of the whitened response with simple matched filtering. MMSE estimation of the whitened response can be accomplished in many cases from this least squares estimate. We compare the approach to conventional maximal length sequences as well as Gold sequence to determine the merits of the approach we further demonstrate the approach for communications at low signal to noise ratios.