Research on AR Adaptive Deck Motion Prediction Technology Based on Forgetting Factor

In this paper, the problem of accurate landing has always been an important problem to be solved in the application process of carrier-based UAV. To this end, the Auto-Regressive (AR) model is used to predict the deck motion process. On the basis of the commonly used system identification method - recursive least square method, the dynamic adjustment of weights for new and old data is realized by introducing a forgetting factor, and the convergence speed of the recursive fitting process is improved by extending the innovation dimension used in the single recurrence. For different sea conditions, the designed deck motion predictor is simulated and verified. Compared with other methods, the results show that the least square method with multi-innovation forgetting factor can control the error of deck motion prediction in a smaller range, and improve the accuracy and success rate of landing.