UACE2017 Proceedings: Optimum space-time filter performance using a realistic noise model of the ambient environment.

• Session:
  Underwater Noise - Modelling and Measurements
• Paper:
  Optimum space-time filter performance using a realistic noise model of the ambient environment.
• Author(s):
  Afolarin Egbewande, Jean-Francois Bousquet
• Abstract:
  To evaluate the reliability of underwater acoustic communication systems, it is typically assumed that the noise at the receiver follows a white Gaussian noise distribution. In this work, a space-time underwater acoustic noise model in the 100-10 kHz frequency band is used to estimate the performance of a multi-hydrophone receiver in a coherent communication system. The objective is to reliably transmit in a shallow water environment a frame of information over a narrow band on the order of 400 Hz and centered around 2.5 kHz. \nA space-time ambient noise model developed in this work is intended to represent shallow water environments, and includes the effects of shipping noise, surface motion, and turbulent flow. The model also includes the effect of correlation as a function of vertical position. The space-time noise coherence in a vertical array of hydrophones is extracted over the bandwidth of interest using a shallow water model. To validate the noise model, it is compared with real measurements of ambient noise taken near the Halifax harbour. A pair of hydrophones were deployed, and separated by 4 meters along the vertical column in a water depth approximately equal to 20 meters. It is found that noise in this region was wind dominated with occasional impulsive noise from marine animals.\nTo evaluate the communication link performance, the signal of interest is modelled in software including the distortion due to propagation. The signal contribution at the input of the receiver is added to the measured noise signal. A previously documented space-time equalizer relying on adaptive decision feedback combined with a phase locked loop is used to evaluate the performance. Because the noise and signal of interest both contain a spatial signature, it is found that detection reliability relies on the ability of the filter to separate the two signals. \n
•   Download the full paper

Contact details

• Contact person:
  Prof Jean-Francois Bousquet
• e-mail:
  
• Affiliation:
  Dalhousie University
• Country:
  Canada