Multisensor Data Fusion Applied to Marine Integrated Navigation Systems

This paper deals with the problem of designing a flexible and accurate integrated navigation system (INS) for marine craft. The proposed INS is based on the integration of a global positioning system (GPS) with a compass and a speed log. After introducing the scopes and functions of the proposed INS, mathematical models of its main components are presented. Then the development of a new multisensor data fusion algorithm for carrying out an accurate estimation of the main state variables is presented. The theoretical background for the sensor fusion is based on the classical Kalman filter theory, which allows to update an a priori position estimate, given by a dead-reckoning system, with the information supplied by a GPS positioning system. Finally the filtering algorithm is extended in the framework of interval analysis and fuzzy set theory in order to improve the reliability and robustness of the estimation algorithm. The validity of the proposed approach is demonstrated by simulation examples applied to a container ship navigating in realistic conditions.