Passive Towed Array Shape Estimation Using Heading and Acoustic Data

This paper addresses the problem of array shape estimation for passive towed sonar systems during platform maneuvers. Directional noise fields due to distant shipping lanes can be exploited as sources of opportunity for online array shape calibration. In this paper, a nonparametric noise field model is used to form field directionality maps for time-varying array shapes to exploit point and spatially spread sources. This formulation requires neither the number nor location of sources in the field to be known or estimated. Using acoustic data, a maximum-likelihood array shape estimate is derived where the shape is modeled as a polynomial in heading. Additionally, a method for fusing the shape estimate with heading sensor data is introduced. Heading sensors may permanently fail or suffer from high levels of noise during turns; thus acoustic data can be used to compensate for malfunctioning heading sensors during turns. The combined estimate is filtered using a dynamical model that is valid for sharp turns and accounts for motion of the array perpendicular to tow heading. Multisource simulations are used to demonstrate the performance of the acoustic-based estimate and robustness of the combined estimate.

Duke Scholars

Author Jeffrey L. Krolik Electrical and Computer Engineering