An airspace complexity unsupervised assessment method

Abstract

The invention relates to a space domain complexity unsupervised evaluation method. The method comprises the following steps: processing sector operation data to obtain an original complexity sample; wherein each sample corresponds to the operation situation of a certain sector in a certain time period; utilizing a kernel principal component analysis method KCPA to non-linearly map the original complexity sample into an infinite-dimension regeneration kernel Hilbert space, then converting the infinite-dimension samples into a low-dimension subspace with maximum complexity assessment informationamount, and extracting m principal components meeting the contribution rate of user requirements from the infinite-dimension samples; secondly, designing a clustering algorithm with multiple adjustable input parameters, configuring the complexity level number, the sample proportion of each complexity level and an initial clustering center by a user according to requirements based on sector operation characteristics to be evaluated, obtaining a clustering result of each original sample level through a clustering experiment, and finally finishing unsupervised evaluation of the airspace complexity.

Description

technical field [0001] The invention belongs to the field of space domain complexity evaluation, and in particular relates to an unsupervised evaluation method of space domain complexity. Background technique [0002] Airspace complexity assessment, as a key means to measure airspace operation situation and controller's work pressure, is the basis of air traffic operation regulation. Due to the large number of influencing factors, the complex coupling and correlation between different factors, and the difficulty in obtaining calibration samples, the accurate assessment of airspace complexity is recognized as a challenging problem in the aviation field. [0003] For the evaluation of air traffic complexity, domestic and foreign scholars have proposed a variety of methods, including Professor Prandini of Politecnico di Milano [[1]Prandini M, Hu J H.A probabilistic approach to air traffic complexity evaluation[C] / / Proceedings of the Joint 48th IEEE Conference onDecision and Co...

AI Technical Summary

Problems solved by technology

Due to the large number of influencing factors, the complex coupling and correlation between different factors, and the difficulty of obtaining calibration samples, the accurate assessment of airspace complexity is recognized as a challenging problem in the aviation field

[0003] For the evaluation of air traffic complexity, domestic and foreign scholars have proposed a variety of methods, including Professor Prandini of Politecnico di Milano [[1]Prandini M, Hu J H.A probabilistic approach to air traffic complexity evaluation[C] / / Proceedings of the Joint 48th IEEE Conference onDecision and Control and 28th Chinese Control Conference, Shanghai, China, December 16-18, 2009.] Proposed flight conflict probability index, Professor Delahaye, National University of Civil Aviation, etc. [[2] Puechmorel S, Delahaye D. New trends in air traffic complexity[C] / / Proceedings of the 2009ENRI International Workshop on ATM / CNS(EIWAC),Tokyo,Japan,March 5-6,2009.] proposed the Lyapunov index index, but the singleness of the calculation angle leads to the measurement of complexity Relatively one-sided

In addition, Professor Gianazza [[3]Gianazza D.Forecasting workload and airspace configuration with neural networks and tree search methods[J].Artificial Intelligence,2010,174:530-549.] proposed to use neural networks to study airspace complexity, but The performance of the machine learning-based spatial complexity evaluation model relies heavily on calibration samples

However, the sample calibration of the airspace complexity requires air traffic control experts to perform manual calibration on the basis of carefully checking the control information, and the acquisition cost is very high.

In addition, due to the dynamic evolution of the static structure and operating mechanism of the airspace, the existing calibration samples will gradually expire. When it is necessary to train a new complexity evaluation model, it is difficult to ensure that the calibration samples are available, resulting in a serious decrease in the reliability of the model.

Images