A new method to abstract the noise model in INS/GNS integrated navigation system

In this paper, we provide a new practical method of abstracting the noise model of an INS/GNSS integrated navigation system from measuring data. After deductively designing the experiment, we carried out tests of the stationary stochastic process on logged GNSS data. The noise model was then fitted to the data and the parameters estimated. The design uses a classical cascaded Kalman filter based on position and velocity integration and applies the theory of stochastic processes, probability statistics, time series analysis and system identification. This method, we think, can be used practically in other integrated navigation systems.