2020/11/18 by Sim Kuan Goh, Narendra Pratap Singh, Goh, Sim Kuan +5
Computer Science · Engineering · #Aerospace and Aviation Technology #Autonomous Vehicle Technology and Safety #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Machine Learning (cs.LG) #Robotics (cs.RO)
paper · pdf · doi:10.48550/arxiv.2011.09335
openalex publication_date 2020/11/18 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Approach and landing accidents have resulted in a significant number of hull\nlosses worldwide. Technologies (e.g., instrument landing system) and procedures\n(e.g., stabilized approach criteria) have been developed to reduce the risks.\nIn this paper, we propose a data-driven method to learn and interpret flight's\napproach and landing parameters to facilitate comprehensible and actionable\ninsights into flight dynamics. Specifically, we develop two variants of tunnel\nGaussian process (TGP) models to elucidate aircraft's approach and landing\ndynamics using advanced surface movement guidance and control system (A-SMGCS)\ndata, which then indicates the stability of flight. TGP hybridizes the\nstrengths of sparse variational Gaussian process and polar Gaussian process to\nlearn from a large amount of data in cylindrical coordinates. We examine TGP\nqualitatively and quantitatively by synthesizing three complex trajectory\ndatasets and compared TGP against existing methods on trajectory learning.\nEmpirically, TGP demonstrates superior modeling performance. When applied to\noperational A-SMGCS data, TGP provides the generative probabilistic description\nof landing dynamics and interpretable tunnel views of approach and landing\nparameters. These probabilistic tunnel models can facilitate the analysis of\nprocedure adherence and augment existing aircrew and air traffic controllers'\ndisplays during the approach and landing procedures, enabling necessary\ncorrective actions.\n