Modeling Surface Multipath Effects in Synthetic Aperture Sonar

Abstract

Synthetic aperture sonar (SAS) imaging algorithms assume a specific ping-to-ping phase relation in the collected data. The line-of-sight signal from a nonmoving object adds coherently from ping to ping in the image reconstruction process while any random multipath reflections or backscatter from the sea surface may add noncoherently, thus improving the image signal-to-clutter ratio (SCR). To move towards understanding just how effective a SAS is at suppressing surface multipath contributions, it is necessary to model the moving surface in a believable way and establish how the sound reflects from the undersurface of the sea. This paper presents a method for simulating the effects of multipath propagation on sonar data and hence evaluating the SCR improvement realized with synthetic aperture processing. This paper first reviews the Pierson-Moskowitz and cos-2s surface-wave spectra, which together account for wind direction, wind speed, and angular spread of the wave propagation direction. From these spectra a statistically appropriate random wave surface is generated which evolves in both time and space. In a first attempt to model the sea-surface multipath problem, a set of impulse responses are generated from this wave surface as it evolves in time increments equal to the pulse repetition period. Two sea-surface scattering mechanisms are used in the simulations described in this paper. In the first, each surface facet reflects as a diffraction-limited radiating aperture and in the second, each facet reflects as an incoherent Lambertian scatterer. These describe two limiting situations: first, the acoustic wavelength is small compared with the roughness of the sea surface; and second, the acoustic wavelength is significant in proportion to the surface roughness. The effect of surface multipath is shown on raw data and also on processed SAS images. The calculation of the SCR as a function of sea state is also demonstrated. The SCR improvement seen with SAS imaging is consistent with the hypothesis that surface multipath signals are fully incoherent from ping to ping.