Data-driven Conflict Detection Enhancement in 3D Airspace with Machine Learning

Abstract

Trajectory prediction with Closest Point of Approach (CPA) concept is a fundamental element of aircraft Conflict Detection (CD) problem. Conventional motion-based CPA prediction model generally assumes that aircraft is flying in straight line with constant speed. But due to environment uncertainties and ground speed changes, this conventional method frequently lacks accuracy in the real world with a high rate of false alarms and missed detections. In this paper, we introduce a novel automated data-driven CD framework with Machine Learning (ML) for 3D CPA prediction in a lookahead time of less than 20 minutes. Firstly, a 3D CPA model with cylindrical norm is proposed as the baseline. Then, data preparation with Mode-S observation data in France is explained, including data collection and data processing, to convert raw Mode-S data to the closeto-reality dataset. Furthermore, feature engineering is applied to build up a feature set with 16 features. Finally, four prevailing ML models are used to predict the time, horizontal distance and vertical distance of CPA in 3D airspace. CD is conducted based on the predicted values. The prediction and CD results show that all proposed ML models outperform the baseline model. Especially, GBM and FFNNs could strongly enhance the performance of CD.

Enhancing network security through machine learning: A study on intrusion detection system using supervised algorithms

“…To figure out which concept should be delivered first, the confusion matrix can be a starting point that lays a solid foundation for both accuracy and the classification report. A confusion matrix is indeed an N*N matrix to check whether the performance of a machine learning model is productive [21]. While concentrating on the letter N, it merely means the number of groups available for investigation.…”

“…After setting up the confusion matrix, the second concept called accuracy is able to be conveyed succinctly. Accuracy is defined as a probability for which how many samples out of the total number of samples are indicated to be correct [20][21][22], whereby the format can be either a percent or decimal based upon the interest of mathematicians or machine learning scientists. Therefore, accuracy under the confusion matrix drawn in Figure 3 can be described as samples that are allocated in True Positive(TP) and True Negative(TN).…”

Preliminary design of an ATC support tool for the implementation of the Ad Hoc Separation Minima concept in an en-route sector.

“…Several studies have delved into complex probabilistic models to grasp trajectory uncertainty [4], [5], [6], [7], [8]. Recently, in the context of Artificial Intelligence (AI), Machine Learning (ML) algorithms have also been employed for CD [9], [10].…”

Machine Learning classification techniques applied to static air traffic conflict detection

“…One of the most extended application may be for trajectory prediction [10]- [12], although it has been used for air traffic flow analysis or safety analysis [13], [14]. Conflict detection is an emerging topic that has been barely tackled based on ML techniques [15], [16]. This work aims to analyse the feasibility of using ML techniques for conflict detection for en-route airspace.…”