Abstract
When the towing vessel was turned, the Directions-of-Arrival (DOA) estimation would be critically degraded if the horizontal towed linear array shape was still assumed to be a straight line. Especially to the coherent signals, which were frequently encountered in shallow water environment due to the multipath propagation, the DOAs could only be estimated by applying the Maximum Likelihood (ML) estimator with a priori information of the array geometry. In this paper, an effective method was proposed to simultaneously estimate the DOAs and the sensor positions of a bended towed array during a turn in the coherent signal environment, which was called the array shape self-calibration problem. Firstly, the spatial signature corresponding to the arbitrary number of incident coherent signals was blindly obtained from the output of the Constant Modulus (CM) array only based on the array sample data. Then the towed array shape calibration problem could be described as a nonlinear optimization problem and solved by Genetic Algorithm (GA). Once the towed array shape has been calibrated, the DOAs of coherent signals would be estimated subsequently by using the ML estimator. Finally, numerical examples were conducted to illustrate the effectiveness of the proposed method.
