Learning Air Traffic as Images: A Deep Convolutional Neural Network for Airspace Operation Complexity Evaluation

Table 2

Summary of previous works in air traffic complexity evaluation.

Category	Related research	Paper
Single model	Difficulty to resolve potential flight conflicts	Lee et al. [12]
	Probability of flight conflicts	Prandini et al. [13]
	Lyapunov exponent of trajectory	Delahaye and Puechmorel [14, 18]
Factor system	Dynamic density method (linear)	Kopardekar and Magyarits [15]
	PCA + BPNN	Gianazza [16]
	Genetic algorithm + AdaBoost	Xiao et al. [17]
	Semisupervised and active learning	Zhu et al. [20]
	Transfer learning	Cao et al. [2]
Image-based factors	Deep convolutional neural network	This paper