Air Traffic Flow Prediction Model Based on Improved Adding-Weighted One-Rank Local-rejion Method

Accurate air traffic flow prediction is an important basis for efficient air traffic control and management. Aiming at the inherent chaotic dynamic characteristics of air traffic flow time series, the chaotic traffic-flow time-series prediction model based on improved adding-weight one-rank local-region prediction method was analyzed herein. Firstly, an improved adding-weight one-rank local-region prediction method was proposed, which involved weighing the evolution of adjacent phase points. Further, the prediction results were corrected by construction of error sequences during the prediction process. Secondly, the chaotic characteristics were verified to exist in four groups of air traffic flow time series at different time scales, using the saturation phenomenon of correlation dimension. Finally, a validation experiment for air traffic flow prediction was carried out using the improved method, after phase space reconstruction of air traffic flow time series. The results show that the prediction accuracy of all four groups is improved, wherein the traffic flow time series with time scale of 10 min has the best precision; the relative error in this case reduces by 29.7%.