Comparison of Kalman and least squares filters for locating autonomous very low frequency acoustic sensors

The Marine Physical Laboratory has designed, fabricated, and taken to sea self-contained, freely drifting acoustic sensors which can measure signal propagation and ambient ocean noise in the 1-20-Hz band for up to 25-hour periods. The deployment of several freely drifting floats forms an array of sensors whose outputs can be combined after the experiment with a beamformer. A Kalman filter and a least-squares estimator have been developed to estimate float positions from travel-time measurements. Computer simulation is used to compare filter performance-under several deployment scenarios. Results show that the Kalman filter performs better than the least-squares filter when the floats are subjected to small-magnitude accelerations between measurements. Neither filter was sensitive to relatively major changes in deployment geometry as long as the sound-speed profile is known exactly.<
>