A comparison between filtering approach and spline approximation method in smoothing flight data

Onboard measurement system, the function of which is to collect, record and process measurement information, performs these tasks by obtaining information from the sensors. However, the sensors are noisy, so it is necessary to combine multiple pieces of information to give a good accuracy to the signals provided by the onboard measurement system. The extended Kalman filter, which is one of the most widely used data fusion methods, is also one of the derivatives of the standard Kalman filter used for non-linear problems. On the other hand, the spline approximation, especially the cubic and Hermitian splines provide a very good and smooth estimate. It is known that the implementation of spline method is simpler than Kalman filtering. This paper describes an empirical comparative analysis of the extended Kalman filter and spline method. The impressive result is that a simple spline approximation in many cases performs better than a sophisticated Kalman filter.