Passive Acoustic Monitoring of Moving Targets by Sparse Underwater Sensor Network
|Title:||Passive Acoustic Monitoring of Moving Targets by Sparse Underwater Sensor Network|
|Series:||Blekinge Institute of Technology Licentiate Dissertion Series|
|Publisher:||Blekinge Institute of Technology|
|Organization:||Blekinge Institute of Technology|
|Department:||School of Engineering - Dept. of Signal Processing (Sektionen för teknik – avd. för signalbehandling)
School of Engineering S- 372 25 Ronneby
+46 455 38 50 00
|Abstract:||Surveillance of sound activities in the sea is of great importance for many applications. A common way to monitor the sea is to register sound with the help of a group of closely placed sensors that form a one-dimensional or two-dimensional array. This is a practical and easy way to monitor sound in large and deep waters where sound spreads in nearly straight lines. In shallow waters, propagation patterns are more complex and also more difficult to predict. Baltic Sea shallow waters also involve large archipelagoes, which make monitoring more difficult. New possibilities to synchronize sensors with the aid of the global positioning system (GPS) facilitate the creation of larger sensor systems. By placing sensors at different positions over the entire water volume under observation, a minimization of the effects from complicated sound spreading is allowed. Near field processing can then also be applied since the sound source is registered by means of several sensors in positions close to the target. These large arrays will have three-dimensional capabilities and they will also provide a much higher resolution than traditional systems.
In this thesis it is shown that signals registered hundreds of meters away from each other can be used to position moving sound within the sensor area. Measurement results are presented from four experiments in shallow waters 20 meters deep, located in the Baltic Sea. The measurements were performed with different setups, at different locations and during different seasons.
The thesis comprises four parts. The first part presents a comparative study of two methods for estimating the time delay between two measurement signals, where the sound consists of noise and dominant tones. Effects from different geometric setups on the position estimate are shown. The result indicates that larger baselines give better positioning capabilities despite a larger uncertainty in the time delay estimation. The second part of this thesis concerns imaging, which means the projection of multiple pairwise correlation results into a two-dimensional plane. Results from a target placed at a distance equal to the length of the sensor baselines are shown. Weights that can be used to give different sensor pairs different importance in the summation image are investigated. In this case, the target was assumed to consist of one sound source. The third part of the thesis also concerns imaging, but presents results from a target placed at a distance smaller than the length of the sensor baselines. In this part, simulation shows that it is possible to position an object in a three-dimensional space using sensors placed in a two-dimensional span. The fourth and final part of this thesis also focuses on imaging and here the results indicate that several sound sources can be resolved for a target.
Signal Processing\Radar and Sonar