A Ship GPS/DR Navigation Technique Using Neural Network

Abstract

A stable and accurate ship navigation system is very important for people navigating in the ocean. Dead reckoning (DR) system is a frequently used navigation system for ships navigating in the ocean. It can provide precise short term navigation data, but the error of DR system can accumulate over time without limitation. GPS can be used for navigation in outside environment. But the accuracy of GPS for civilian use is still big. In this paper a cheap single GPS receiver is used for the navigation for ships. This paper proposed a new Kalman filter based GPS/DR data fusion method using neural network. This method was designed based on the characteristic of the GPS receiver. By using this data fusion method the cheap single GPS receiver can cooperate with DR system to provide precise navigation information. Simulation is conducted to validate the proposed data fusion method.KeywordsKalman filterGPSdead reckoningdata fusion methodBP neural network