Multidimensional geomagnetic matching fusion algorithm based on projection pursuit

The algorithm of single characteristic parameter geomagnetic matching will be influenced by the slow changing of characteristic parameter easily, and has reduced the orientation precision of the matching. A method of multidimensional geomagnetic matching fusion algorithm is presented, which aims at the territorial uncertainty problem of the single parameter matching method. This method, which based on the characteristic parameter´s projection pursuit classification model of multidimensional geomagnetic matching, has been calculated by the coevolutionary algorithm. The finally position is confirmed by specific weight (reliability) of multidimensional geomagnetic parameter, which gained by the character of sample data. The algorithm has avoided the territorial uncertainty problem. Experimental results show that the error of adaptive parameter matching system is much lesser then the single parameter matching system. Simulations verify the effectiveness of the proposed method.