A sequential filtering algorithm for range estimation of a moving ship based on striation geometric features

Abstract

There exist regular striation pattern in the broad-band noise field excited by a moving ship in shallow waters, whose time (range) frequency structure can be interpreted by the waveguide invariant theory. The striation characteristics have been exploited for underwater acoustic inversion problems, e.g., sediment geoacoustic parameter characterization. In this paper, the geometrical features of specific striations are observed for continuously ship ranging by a sequential algorithm. Comparing to matched-field processing techniques, this approach also has no requirements prior knowledge of environmental properties and burden of forward sound field calculation, which is similar to virtual receiver approach. The technique is tested on the striation fragments extracted from the ship noise data collected in Dalian 2008, and the outputs are in high agreement with in-situ GPS measurements. The results from independent runs with different initial values also suggest that this approach is robust to initial conditions in the sense of all runs can converge to the true values rapidly.