Experimentations and Sonar Development for Buried Objects Detection and Classification Mélanie FOULON 1 , Maud AMATE 1 , Nicolas BURLET 2 , Paul PENVEN 2 , Pierre CERVENKA 3 , Jacques MARCHAL 3 1 DGA Techniques Navales, Avenue de la Tour Royale BP 40915 83050 TOULON Cedex, FRANCE 2 Thales Underwater Systems, Route de Ste Anne du Portzic CS 43814 29238 BREST Cedex 3, FRANCE 3 Institut Jean Le Rond d’Alembert, UPMC, 2, place de la gare de ceinture 78210 Saint-Cyr-l’École, FRANCE Abstract In order to be able to detect, localize and classify buried objects in the sediment, DGA Naval Systems decided to study and realize dedicated sonar. The principle of this low frequencies sonar is to use the advantage of parametric transmission and synthetic aperture techniques in reception. Experiments in tank with buried objects in sand have been conducted. During these experiments, several objects with diverse sizes were buried in two types of sediment and different signals have been transmitted. The first exploitations show that it is possible to detect buried objects with such techniques. Using the results of these experimentations, a sonar has been built. It comprises a parametric transmission antenna and two reception antennas: one for high frequencies (100 kHz) and one for low frequencies (20kHz). The performances of this sonar have been characterized in tank. This article will present experimentations, the sonar and the results. The next step is to organize sea experimentations to test sonar performances on operating conditions. Introduction Low frequency, large bandwidth and narrow beams are required to respond to the threat of buried objects in order to respectively have a correct sediment penetration, have a good resolution and limit reverberation. The combination of parametric transmission and synthetic aperture processing is a suitable approach for designing a buried object detection sonar because this two methods in transmission and in reception employ a smaller antenna than with conventional using. This article tackles the subject of an experimental sonar based on a sequential multibeam sonar prototype associated with a synthetic aperture in ventral configuration. However, this drawback of the technique is the coverage rate limitation. To increase the time coverage, chosen solution is the optimised simultaneous generation of several parametric beams in different directions. The article presents the principle of parametric transmission and SAS processing. Then trials, conducted in April 2008 by GESMA, Brest, are described. Finally the first images obtained after processing are shown. DGA contracted Paris VI University for studies about parametric transmission theory and SAS processing theory. DGA contracted Thales Underwater Systems for processing. Parametric and SAS theory Parametric transmission is obtained by creating two overlapping primary beams at frequencies F 1 and F 2 . The non- linear interaction in the water of these waves creates other waves at frequencies which are linear combinations of F 1 and F 2 . The wave of interest is at the difference frequency F BF = F 2 - F 1 . The main interest of the parametric transmission is the high directivity that can be obtained with an antenna whose size is not very large compared to the wavelength. The other benefits are the large bandwidth potentially available and the low level of the side-lobes in the directivity pattern. However, a significant drawback is the poor efficiency of the non-linear conversion. The Westervelt model [1] draws a simple picture of the parametric transmission principle. The primary waves are modeled by collimated beams. They make a source line at the difference frequency that behaves as an end fire array. Such geometry creates a beam pattern whose farfield directivity is proportional to the square root of the array length which is only limited by the attenuation of the primary waves. Such a model is convenient to figure the principle of