LAUSR

The usual approach to calculating the time domain acoustic response of an ocean waveguide is to Fourier synthesize it from a regularly sampled collection of frequency domain solutions. Depending on the details of the waveguide and the excitation signal, the resulting time series may be very sparse with significant time intervals of relative quiet between energetic multipath arrivals. This observed sparsity in the time domain suggests a computational advantage to synthesizing the time series by solving a convex optimization problem involving a much smaller set of solutions computed at random frequencies. The efficiency of the compressed sensing approach will depend on the sparsity of the time domain solution; the greater the sparsity in the time domain signal the greater the expected advantage of the compressed sensing approach. In this talk we examine the compressed sensing method of time series synthesis for a sequence of waveguides that range from extremely simple and ideal waveguides where the sparsity...