Signature Modeling for Acoustic Trainer Synthesis

As shipboard acoustic sensor systems have increased in complexity, embedded trainers have gained acceptance as a means of providing continuous unalerted operator training. An essential element of these trainers is a contact signature synthesizer which generates simulated sensor outputs equivalent to those produced by a received maneuvering contact. In this paper, we give an overall contact signature model broken down into source components. For each of these components, a mathematical model which relates the source-receiver geometry and the contact maneuver state to the component spectral and temporal characteristics is given. The synthesis of both tonal and broad-band components over a frequency range of 0 to 2 kHz was accomplished at simulated contact ranges of 1 to 10 kyd (0.9 to 9 kin). Tones with independent frequency jitter were synthesized. The bandwidth of the jitter could be programmed to vary directly with contact speed. A time-varying bias or wander was induced to model slow drifts in the rate of mechanical machinery rotations. Representative synthesis results are displayed and other acoustic applications are suggested.