Blind identification of medium with multi-layer structure

There are a category of problems about internal structure identification for a multi-layer medium according to reflection wave measurement on the surface. The mathematical model is that the output signal, i.e., the reflected wave series, is a convolution of the input signal, i.e., the exploration wave and the sequence of reflectivity at the interface positions between every two layers, which describes the multi-layer structure. The difficulty for this problem is that the reflectivity sequence has to be identified with the input signal unknown, since the exploration wave is seldom obtained in many cases. The presented methods are all based on some additional assumptions and/or other information, so they can´t guarantee to solve this problem in any case. To overcome this drawback, an essentially new approach is proposed based on blind identification theory. According to our method, reflection wave measurements are performed twice at every exploration point by different exploration waves so that the two exploration waves can be estimated from the two output signals, and then the reflectivity sequence can be obtained by some deterministic deconvolution algorithm. An example for seismic prospecting is given, which demonstrates the effectiveness of this new method