The Study On The Application Of Passive Location In The Underwater Target Tracking

With the development of martial electron technology, especially the hydro-acoustic warfare technology, there are higher requests for the location, tracking and recognition of underwater target. The traditionary active sonar system does not meet the demand of practical battlefield. Passive sonar location system, which is of long operation distance, good concealment and electronic jamming, has always been given deeply recepted by the army.Underwater detect technology is not only important to the martial application, but also to civil application. It has been applied directly or indirectly to ocean examen, underwater resource development, fish finder, collision avoidance, navigation, the monitoring of ocean-atmosphere circulation and so on.The application of passive location in the underwater target tracking is discussed in this paper. First, combined K-Means Cluster and kernel learning, a kernel-based K-Means Cluster is used in order to overcome the shortcoming of K-Means Cluster and solve the data association of passive location. Second, the bearing location is improved based on the studies on the bearing location, time difference location and bearing time difference location. To reduce the angle error and improve the locating precision, the method of least squares is applied to the bearing location. The experiment shows that it has good result.Lastly, using the idea that combine the bearing location and time difference location together, a new mechanism of bearing time difference location is put forward, which put multi-passive-sonar on a circle and a sonar having high bear precision on the center of the circle, and promote the poison of target according to the geometry relations and the law of cosine. The precision of this method is analyzed, and the experiment shows that this new method is effective and feasible. At the same time, data fusion is applied to this new location. To attenuate its randomness, we take the noisy data that multi-passive-sonar locates as a shock disturbed sequence and use the buffering operator upon it. Finally, the technique of data fusion is applied upon its result to further improve the tracking precision of the system. Our experiment shows that the solution is effective and feasible.