Implementation of a gps/ins tightly-coupled system applied in urban environments

Abstract

This paper presents the implementation of a GPS/INS tightly-coupled system applied in urban environments. In this integration technique, the information of satellite distance and Doppler frequency is extracted and calculated for building the measurement model in the integration system. Accleration, gyroscope and magnetic value from INS are used for the process model. An 23-state extended Kalman filter is formulated to estimate position, velocity, bias, scale factor, and clock error. The experiment is carried out outdoor in urban environments where GPS signals are randomly lost in order to evaluate the system accuracy. The experimental results have shown the GPS/INS tightlycoupled integration gives better performance that the GPS/INS loosely-coupled integration.