An aircraft path prediction method based on depth memory network

Abstract

The invention discloses an aircraft track prediction method based on a depth memory network. Firstly, the method extracts the historical aircraft track data and preprocesses the data to form an aircraft track information data sequence under a time sequence. On this basis, the input and output sample characteristics are constructed, and the samples are standardized to generate dimensionless training data set. Then, the attention mechanism model is integrated into the long-term and short-term memory network structure, and the structural parameters are initialized to construct the track prediction neural network. Finally, the track prediction network is trained and optimized, and the track prediction network model with high prediction accuracy is obtained.

Description

technical field [0001] The invention relates to an aircraft trajectory prediction method based on a deep memory network, which belongs to the technical field of civil aviation. Background technique [0002] In recent years, the air transport industry has continued to develop rapidly, and the contradiction between limited airspace resources and increasing air traffic flow has deepened day by day, which has intensified potential conflicts between aircraft, increased the load on controllers, and frequent problems such as airspace congestion and flight delays. How to use effective air traffic management methods to fine-tune the allocation of airspace resources, alleviate flight delays, and effectively detect and resolve conflicts are the main problems facing the development of air traffic. [0003] Track prediction technology is one of the key technologies of air traffic management. Effective and accurate prediction of the four-dimensional track of aircraft is the core of proble...

AI Technical Summary

Problems solved by technology

(2) No parameter estimation

[0005] There are currently the following problems in the prior art: (1) The prediction effect is unstable and the universality is insufficient

Due to the strong randomness of the aircraft track, the single parameter estimation based on the kinematic model will cause a large difference in the prediction accuracy of objects with different motion characteristics, and the universality is insufficient; (2) Lack of learning ability

The characteristics of the predicted track obtained by parameter-free estimation are limited to the features of the existing track, and the prediction performance is relatively related to the clustering performance of the original track data itself, and the feature prediction accuracy for multiple tracks is poor. Lack of learning improvement ability; (3) Low learning efficiency

According to traditional machine learning methods, a large sample size is the guarantee of prediction accuracy, and in the process of learning and training a large amount of data, learning efficiency, that is, learning rate and learning quality, greatly affects the final prediction quality

In four-dimensional aviation data, the degree of correlation between the hidden states of different feature information is different, while the traditional machine learning method learns the information of each feature according to the same degree of hidden state correlation, and assigns unnecessary calculations to some feature information. resources, resulting in low learning efficiency

Images