Multipath characterization of hydroacoustic signals by autocoherence estimation

In shallow waters with pronounced multipath the effects of wave propagation can have far reaching consequences for the ability to perform sonar signal processing tasks such as bearing estimation. The authors discuss the possibilities of using a new measure of coherence, called autocoherence, based on modulation properties of the signals to quantify the multipath effects. More precise knowledge of multipath effects can be used to optimize the preprocessing schemes necessary in complex scenarios where standard estimation methods will encounter problems. To study the effects of multipath the authors have used hydroacoustic narrow-band continous-wave signals from the Baltic Sea generated by a small motor boat approaching and passing an array of sensors. In the autocoherence analysis the authors employed signals in a frequency band of 200 to 750 Hz generated at source-receiver distances ranging from 0 to 800 m. The authors used bicoherence analysis for the identification of frequencies and their interrelation with engine and propeller parameters. The results indicate a lower value of autocoherence at shorter source-receiver distances, i.e. less than 180 m. At longer distances, i.e. 300 to 600 m, the autocoherence value is significantly larger. This is consistent with ray-based propagation theory at the studied frequencies. Autocoherence estimation seems therefore to be a possible method to characterize the multipath effects in shallow water and act as an input to sonar systems to aid the decision of spatial filtering/smoothing method and the choice of parameters