The article presents the putting into practice of a neural state estimator for navigational measurements. One of the basic problems of sea navigation is considered, which is the statistic working out of navigational measurements with the purpose of determining the estimated vector of the vessel's movement and position.

In this paper there are presented results of applying the methods of the time series analysis to the problem of recognizing small boats. It has been showed that the acoustic signals of the boats can be classified by means of clustering algorithms.

The problems of identification of objects have been strictly bound with sea navigation for a long time. The article presents a method of object identification by means of artificial neural networks. The classification of objects is effected on the basis of a radio signal. A comparison between classical methods and neural methods has been made. The results of research performed have been presented on real data, and a detailed analysis of the probabilistic network and its identification.