A new method is presented for passively tracking a submarine with respect to N omnidirectional sonabuoys by utilizing the noise perturbed Doppler shifts in one or more frequencies emitted by the target. In contrast to contemporary approaches to this problem such as Kalman filtering techniques, the approach adopted here requires no apriori assumptions of a Gaussian noise field or straight‐line motion and constant target speed. An algorithm based on accelerated‐gradient optimization techniques is used to track the target by outputting a position and velocity estimate for every time period τ that the FFT processor forms an output. Thus, the algorithm is fast enough to follow the target in real time with a nonincreasing lag. Computer simulation shows this method holds much promise in competing with other alternatives, particularly in following maneuvering targets such as a sinusoid trajectory through a buoy field.

This content is only available via PDF.
You do not currently have access to this content.