2023_programme: Improving Sonar Simulation with a Bistatic Active Target Model

Back


  • Session: 24. Sonar Technology
    Organiser(s): -
  • Lecture: Improving Sonar Simulation with a Bistatic Active Target Model
    Paper ID: 1904
    Author(s): N/A
    Presenter: Bossér Lennart
    Abstract: At the 2017 UACE conference we gave examples on how active sonar performance predictions could be improved by simulating the sonar response with a time domain ray trace model coupled to a target echo model (TEA, Target Echo Analyzer). Using the sonar equation with wave propagation calculations in the frequency domain for a broad band active sonar generally predicts detection distances much longer than those actually experienced in the field by sonar operators. Our solution to this problem was to simulate the sonar detection problem with a ray trace model working in the time domain, coupled to the target echo model. We then observed a loss in detection performance in accordance with field trial results. We attribute this loss to time spread due to multipath transmission. A limitation was that TEA could only handle monostatic reflections. Thus, for rays where the reflected angle was different from the incoming angle (i.e., the bistatic case) we had to replace the two angles with the bisectrix between them.\nThe new target echo model, GRATIS (Graphics Accelerated Target-strength Integration Scheme), demonstrates a method to calculate the target strength for any acoustically rigid body. By using standard libraries from computer graphics, most of the geometric and acoustic processing is performed by the GPU, thus giving fast execution times. Bistatic acoustics is supported, but, since the Kirchhoff approximation is used, good results are only expected within limited bistatic angles. Here we propose a method to calculate the impulse response for the body in a similar fashion to make the target model useful with a time-domain propagation software.\nExamples are given to demonstrate how sonar signal simulations can improve sonar performance predictions.
      Download the full paper
  • Corresponding author: Mr Lennart Bossér
    Affiliation: Bossér Utveckling AB
    Country: Sweden
    e-mail:

Back