Mult-path Signal Processing Technology Based On Homomorphic Filtering

Borderline reflection and the non-uniformity of medium make the underwater acousticchannel have multipath interference. So how to take measure to restrain the multipathinterference, has been a hot issue in underwater acoustic field ever since. The paper focuseson the the Homomorphic Filtering theory for multipath signal processing.First, this paper elaborates the basic theory and current status of the HomomorphicFiltering, studies the cause of the phase winding on the cepstrum operation and itsoptimization algorithm. The impact of the FFT on cepstrum algorithm is also discussed whenit replaces the Z-transform for practical use.The pole-zero positions in minimum phase sequence are analyzed for their effect on thecepstrum based on the property of the signal cepstrum.Studies the characteristic of several kinds of tipical signal, such as CW, LFM and HFM, discussed the signal form effect on theircepstrum feature through simulation and comparison. The cepstrum expression for theimpulsive response of underwater acoustic channel is deduced and several channels withdifferent sound velocity distribution are studied through simulation.Then, the paper does research on the typical filters design for cepstrum signal andchannel separation, including low pass filter, comb filter and peak-search filter proposed inthe paper,and makes comparison among these filters on their advantages and disadvantages.The paper also proposes a original signal constructiong method called dual signal contrast,the simulation and test verified their validity.Finally, the paper employs relativity between reconstruction signal and original signal asreference standard, and analyzes different parameter effect on signal reconstruction, and theparameters involve relative distance, depth, signal-noise ratio and sending signal form; andthen the applicable condition of homomorphic filtering for signal and channel separation isconcluded.