美国数学建模竞赛特等奖论文——1996 A O Detection of a Silent Submarine【数学建模】.pdfVIP

Detection of a Silent Submarine 241 Detection of a Silent Submarine from Ambient Noise Field Fluctuations Andrew R. Frey Joseph R. Gagnon J. Hunter Tart Wake Forest University Winston-Salem, NC 27109 Advisor: Stephen B. Robinson Summary We developed a method for detecting intrinsically silent submarines in the ocean by measuring only the fluctuations in the ambient noise field. This method allows us to calculate the position, velocity, and approximate size of a submarine. Our model relies on measuring the noise field at four different listening stations, with each station composed of four microphones a relatively small distance apart. We calculate an amplitude spectrum of the noise at each micro- phone using a Fourier analysis and compare this spectrum to the previously measured baseline spectrum for ambient noise. The difference between these spectra is the noise reflected from the submarine. We use the four microphones at a particular station to measure the gradient of the peak amplitude from the submarine noise spectrum. Because amplitude varies inversely with distance from the submarine, we can compute the sub- marine’s location from the amplitude and the gradient at each listening station. The approximate size, in terms of the radius of a similarly sized sphere, follows from the distance and peak amplitude. Our comparison of the frequencies of the peak amplitudes of the submarine and ambient noise spectra provides a measure of the Doppler shift at each listening station caused by the submarine’s motion. The Doppler shift gives us a component of the submarine’s velocity in the direction of each station. We select a basis from among unit vectors in these four directions and convert the submarine’s velocity into standard Cartesian components. We wrote a Fortran program to implement our algorithm. Our simulations show that