간섭패턴매칭을 이용한 수중 광대역 음원 위치 추정

Abstract

The localization of acoustic source in a shallow water is an important problem in a passive sonar system. For several decades, it has been considered many times over the past years and in order to improve the performance many attempts has been achieved.

In order to localize the source in passive sonar system, several methods for a long range source can be categorized into two principles: one is based on the curved-wave-front beamforming, which issued in an array of sonar system. It can not estimate with accuracy because it is not consider multi-path effects in an ocean waveguide. The other is based on the waveguide invariant theory which is recently introduced and researched various areas.

The waveguide invariant theory provides a descriptor of the broadband interference pattern on the ocean. The invariant parameter called β which is the slope of the interference pattern is useful for describing the acoustic interference pattern in a waveguide. The range of a source can sometimes also be estimated by the much simpler waveguide invariant method. However the waveguide invariant method requires knowledge of certain ‘invariant’ parameter, which unfortunately often vary significantly with ocean sound speed structure.

Recently many methods are proposed using the waveguide invariant theory in passive application and showed the enhanced performance in various area such as MFP based on acoustic propagation model in multi-path environment and a source ranging. But they are still dependent on the β and affect by ocean environment mismatch. So it is necessary to localize the source that independent of the β without the ocean environment information.

In this thesis, a source localization using two sensors in an ocean waveguide is proposed. The principle employed for a source localization is based on the IPM(Interference Pattern Matching) algorithm, which uses the unique characteristic of the ocean waveguide.

The interference pattern which seen in the sensor spectrogram collected from the moving ship-radiated noise arises from the mutual interference between modes reflected by the surface and the bottom. The interference pattern is directly proportional to the source position by the waveguide invariant theory. If two sensors are used to a source localization in identical acoustic propagation environment, the β has identically effect on the each sensor. Because it is possible to detect the target without regard for β. The proposed IPM algorithm estimates the ratio of each interference pattern from two spectrograms. Therefore the interference pattern ratio between two sensors is the source range ratio. The proposed algorithm requires no knowledge of the ocean environment and independents of the β, although the β varies on the ocean.

Finally, the source localization calculates simply the intersection point of two equations which utilized the Appolonius's circle and the time difference. The Appolonius's circle which defined as the locus of a point whose distance from a fixed point is derived the result from the IPM method and estimates the locus of a source using the circle equation. The time difference is derived the TDOA(Time Difference of Arrival) method and estimates the locus of a source using a hyperbolic equation.

We performed simulation of 3 scenarios to test the IPM algorithm and localization method proposed in this paper, and then practiced error analysis of the results. And finally we tested performance of a real-data collected during MAPLE-05 experiment applied to the proposed algorithm. In simulation result, the IPM algorithm seemed to have excellent performance whose mean error is within 5%, also the localization performance has similar whose mean error within 10%. In the experimental result, proposed algorithm is even available in real ocean because the estimated position of target shows errors of within 10% which is similar to simulation results.

The suggested method can be applied to adapt sufficiently to the field of active sonar and passive sonar using the property of broadband signal.