Adaptive Technique for Eliminating Motion Contamination from Towed Temperature Arrays

Abstract

An Ocean Slope-Displacement Estimator (OSDE) was previously constructed, which estimates the slopes of isotherm displacements from towed thermistor array data. This estimator minimized the problems associated with finestructure and was effective in reducing the contamination caused by vertical sensor motion in these displacement estimates when the rms sensor displacements were approximately 5 to 10 cm. However, when the motions of the sensor were as high as 50 cm, this processor was able to remove only about 90% of the motion contamination in the data, and the remaining contamination was high enough to severely limit the usefulness of the data. Although a mechanical motion isolation system in combination with the processor would permit its continued use under these conditions, such a system is complicated and expensive, and further improvements in the processing technique are worthwhile. An adaptive filter that is based upon a least mean squares approach used in electronic noise cancellation schemes has been developed and incorporated into the OSDE processor. This filter removes all observable effects of sensor motion from the displacement data even when rms sensor motions am as high as 50 cm. The effective noise reduction with this filter is about 16 dB which is a 7 dB improvement over the method of point-by-point subtraction of the motion displacements that was used in the previous OSDE. This improved performance provides the capability to operate towed thermistor arrays without motion compensation devices under most conditions encountered at sea.