A Gaussian process based decision support tool for air traffic management

Eerland, Willem; Box, Simon; Fangohr, Hans; Sóbester, András

doi:10.2514/6.2017-4264

Cited by 3 publications

(10 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A few studies put forwards to develop supporting tools to aid procedure design and landing monitoring in real-time. Prominent amongst them uses generative models, such as Gaussian mixture model [7] and Gaussian process (GP) model [23], to generate trajectories similar to historical data. The generated trajectories provide a data-driven approach to examine new procedure design, compared to non-generative methods based on physical equations of motion [34,35], trajectory clustering methods [36,37].…”

Section: Literature Reviewmentioning

confidence: 99%

“…The aforementioned methods performed time re-sampling, realignment of time 𝑡 = 0, dynamic time warping [22,23] and trajectory reconstruction [7] to handle trajectory with variable size or irregular time interval. While these pre-processing steps simplify the analysis, it can distort the temporal information and introduce unnecessary noise.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Besides, most of the analyses are discrete, except GP-based methods [22,23] that model trajectory continuously and provide a "tube" of historical landing profile. GP provides a principal, practical and probabilistic approach to data modeling.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The classical form of GP has been adopted in [22,23] to model and visualize 3D "tube" of historical aircraft trajectories in Dallas Fort Worth and Denver airport. However, GP suffers from high computational complexity in modeling a large amount of data, which hinders real-time application.…”

Section: Literature Reviewmentioning

confidence: 99%

“…The issue has attracted attention from the academic community and several industries to develop interpretable machine learning models [21] used in safety-critical environments and settings. In aviation, several interpretable machine learning methods have been used to address issues concerning approach and landing safety [7,17,22,23].…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Tunnel Gaussian Process Model for Learning Interpretable Flight’s Landing Parameters

Goh

Lim

Alam

et al. 2021

Journal of Guidance, Control, and Dynamics

View full text Add to dashboard Cite

Approach and landing accidents have resulted in a significant number of hull losses worldwide. Technologies (e.g., instrument landing system) and procedures (e.g., stabilized approach criteria) have been developed to reduce the risks. In this paper, we propose a data-driven method to learn and interpret flight's approach and landing parameters to facilitate comprehensible and actionable insights into flight dynamics. Specifically, we develop two variants of tunnel Gaussian process (TGP) models to elucidate aircraft's approach and landing dynamics using advanced surface movement guidance and control system (A-SMGCS) data, which then indicates the stability of flight. TGP hybridizes the strengths of sparse variational Gaussian process and polar Gaussian process to learn from a large amount of data in cylindrical coordinates. We examine TGP qualitatively and quantitatively by synthesizing three complex trajectory datasets and compared TGP against existing methods on trajectory learning. Empirically, TGP demonstrates superior modeling performance. When applied to operational A-SMGCS data, TGP provides the generative probabilistic description of landing dynamics and interpretable tunnel views of approach and landing parameters. These probabilistic tunnel models can facilitate the analysis of procedure adherence and augment existing aircrew and air traffic controllers' displays during the approach and landing procedures, enabling necessary corrective actions. Nomenclature 𝐺𝑃 = Gaussian process models 𝐾 = kernel function defined by 𝐺𝑃

show abstract

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Section: Literature Reviewmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Tunnel Gaussian Process Model for Learning Interpretable Flight’s Landing Parameters

Goh

Lim

Alam

et al. 2021

Journal of Guidance, Control, and Dynamics

View full text Add to dashboard Cite

show abstract

Real-time Unstable Approach Detection Using Sparse Variational Gaussian Process

Singh

Goh

Alam

2020

2020 International Conference on Artificial Intelligence and Data Analytics for Air Transportation (AIDA-At)

View full text Add to dashboard Cite

Worldwide, Air Navigation Service Providers (ANSP) are striving to exceed the desired safety levels. The Terminal Manoeuvre Area (TMA) is one of the most safetycritical areas in ATM as it encompasses the most critical phase of flight, i.e., departure and landing. An aircraft, during the final approach phase, is required to remain in a stable configuration and prevent any undesired state such as an unstable approach, which may subsequently lead to incidents/accidents such as Go-Around, Runway Excursions, etc. In this paper, we propose a data-driven framework to model the aircraft 4D trajectories in the final approach phase by adopting sparse variational Gaussian process (SVGP) model. The model is trained to learn the aircraft landing dynamics from Advanced Surface Movement Guidance and Control System (A-SMGCS) data, during the final approach phase. We experimentally demonstrate that SVGP provides an interpretable probabilistic bound of aircraft parameters that can quantify deviation and perform real-time anomaly detection. The findings of this work can increase situational awareness of the air traffic controller and has implications for the design of a new approach procedure in complex runway configurations such as parallel approach.

show abstract