Adaptive unscented incremental filter method

The adaptive unscented incremental filter (AUIF) model was put forward, in which its concept, model, basis equations and the recursive calculative steps were established. Classical filters did not research the system errors of measurement equations. Due to environmental factors and the instability of measurement equipments, it is difficult to accurately obtain the measurement equations that are not verified and calibrated. So the measurement data have unknown time-varying system errors in actual engineering in the actual environment (such as deep space exploration). The model equations and the noise characteristics have many uncertainties. The uncertainty will lead to greater Kalman filtering error and indeed diverges. The presented AUIF can successfully eliminate the measurement equation system errors. The method can estimate statistical characteristics, adjust gain matrix in real time and greatly improve the filtering accuracy. The method is simple to calculate and easy to apply in engineering.