Figure 1 - uploaded by Duc-Thinh Pham
Content may be subject to copyright.
Source publication
Contexts in source publication
Context 1
... Interpolating sample points between reference points to complete trajectories. The generated trajectories are illustrated in Figure 1. ...
Context 2
... month ADS-B data (December 2016) for South-East Asian region is obtained from an Aviation data company known as FlightAware. For this research purpose, we have identified Sector 2E (Figure 1), an En-route area within Kuala Lumpur FIR, managed by Singapore ACC, for providing air traffic service from FL120 to FL360 inclusive. It takes about 5 minutes in average for a typical flight to cross sector. ...
Similar publications
Citations
... Thus, our generative model is, in fact, a set of three GP models. The work in [21] develops a Homoscedastic GP model to capture the consistent variance in data. In our work, the variations in longitude, latitude, or altitude of aircraft positions vary with time. ...
... The historical data will be utilized for training and evaluating the proposed approach. Its performance will be compared with a homoscedastic GP approach, adapted from [21], for accessing the advantages of our approach in modeling the heteroscedastic noise in real trajectory data. ...
... For comparison, a baseline probabilistic trajectory prediction model using classical homoscedastic GP adapted from [21] is implemented, trained, and tested with the same set of data. The performances of those models are accessed both qualitatively (e.g., visualizing and observing the variation of the predicted trajectories from both approaches) and quantitatively (e.g., com-paring the Kullback-Leibler (KL) divergence obtained from both approaches). ...
Conflict detection plays a crucial role in ensuring flight safety and efficiency and is a critical component of an air traffic control system. Despite the availability of tools to support air traffic controllers in identifying potential conflicts, their quality, and accuracy remain limited due to the challenge of accurately accounting for uncertainty when predicting flight trajectories. To tackle this issue, researchers have explored various studies focused on using probabilistic techniques to model aircraft dynamics and trajectory uncertainty. However, these approaches share several common shortcomings, including their assumptions about uncertainty distributions and the high computational costs of detecting and calculating the risk of conflicts. In response to these challenges, we propose a data-driven approach combining a multi-output generative model with a Bayesian Optimization algorithm to effectively model the uncertainty of aircraft trajectories and rapidly identify the probability of a conflict. Our approach employs the Heteroscedastic Gaussian Process to capture complex trajectory patterns and uncertainty from historical data directly. The proposed predictive model can effectively capture heteroscedastic noise from real data, leading to improved predictions. It achieves Kullback-Leibler divergence around 1 to 1.3 for all dimensions which reduces by > 45% for latitude, > 24% for longitude, and 4% for altitude compared to the classical homoscedastic GP approach. The method also boasts high-performance predictions for 4D trajectories including descending, climbing, and en-route phases. To pinpoint when two aircraft are most likely to experience a conflict, the Bayesian Optimization algorithm is adopted, which shows good performance in terms of computational efficiency and flexibility for probabilistic conflict detection. The proposed model achieves percentage error < 0.25% in estimating the conflict probability with computational cost < 14s. By addressing the challenges of uncertainty and computational complexity, our method demonstrates great potential to enhance flight safety and efficiency.
... IEEE, 2019. Virginia, USA [61]. ...
The increasing in traffic demand has strained air traffic control system and controllers which lead to the need of novel and efficient conflict detection and resolution advisory. In the scope of this thesis, we concentrate on studying challenges in conflict detection and resolution by using machine learning approaches. We have attempted to learn and predict controller behaviors from data using Random Forest. We also propose a novel approach for probabilistic conflict detection by using Heteroscedastic Gaussian Process as predictive models and Bayesian Optimization for probabilistic conflict detection algorithm. Finally, we propose an artificial intelligent agent that is capable of resolving conflicts, in the presence of traffic and uncertainty. The conflict resolution task is formulated as a decision-making problem in large and complex action space, which is applicable for employing reinforcement learning algorithm. Our work includes the development of a learning environment, scenario state representation, reward function, and learning algorithm. Machine learning methods have showed their advantages and potential in conflict detection and resolution related challenges. However, more studies would be conducted to improve their performances such as airspace network representation, multi-agent reinforcement learning or controller's strategy reconstruction from data.
