Coastal traffic density has increased recently due to an expansion of the international sea trade, a growing of the world economy and an enlargement of the demand to the fishery resource. The steady increase of marine traffic has often resulted in the cause of marine accidents with the exorbitant economic loss and the environmental pollution at the sea. In the last 10 years, the ship's collision has recorded about 21.8% of the total marine accidents. More over the most of these collisions have caused by navigator's negligence of watch keeping during works. In particular, about 25% of the collisions have happened in navigation under the restricted visibility condition like a foggy weather. In the foggy condition, navigator must hear the whistle, bell and/or the siren of lighthouses or ships. The whistle blast is a very important information for safety sailing. Even though the navigator can get the brief informations like the direction and range of a sound source, it is not enough to make a suitable decision for its relative position. More over the duty officer cannot hear outside sound signal on the vessel which bridge is totally enclosed, this causes a significant problem. Therefore IMO(International Maritime Organization) has adopted the 'Sound Reception System' as new navigation equipment which can receive the whistle sound signal and indicate the approximate direction of incoming signal.
In this study, it is designed to develop a sound source tracking system that is able to measure the range and relative bearing of sound source by utilizing the whistle blast when other vessels are not identified on the radar screen under the foggy weather.
Three microphones receive the whistle sound in the proposed system. When one microphone is installed as a reference, the others have received the sound signal with the arriving time delay due to the distance difference between microphones. It is able to measure the relative position with range and bearing by considering these amount of time delays between microphones and the trigonometric relation of microphones array and source.
This paper focuses on the principal of measuring the time delays and three calculation methods for finding them between three received digital data sequences. These are the cross-correlation analysis, fast cross- correlation analysis and subtracting method. Two algorithms which approximates the sound source's bearing and distance were also su- -ggested, one could be approximated the gradient of hyperbolic asymptote as a sound source's relative bearing, the other could be approximated under the condition of a long range source relative to the microphones interval as a source's range.
In the system three digital filters were designed to develop the optimal system which could find the time delays between each microphones with high accuracy and applied to the digital signal processing part. These are the non-recursive low pass filter using kaiser window method, recursive inverse chebyshev low pass filter and recursive inverse chebyshev high pass filter.
As a result, we have found that the accuracy of measurements were differentiated by the methods what kind of digital filter were adopted. And we have confirmed the facts that the digital signal processing method using IIR HPF together with FIR LPF was suitable for the positioning of source, where the approximation methods could measure the bearing and distance with higher accuracy than the method using trigonometric relation could do.
We have recognized that sound source tracking system is possible to the sea field with improvement of position error. When source relative bearing was within -50˚~+50˚, the measured bearing was accurate. This system shall be a very helpful to make a safety navigation and reduce the marine accidents.