A turnaround control system to automatically detect and monitor the timestamps of ground service actions in airports: A deep learning and computervision based approach br

Yildiz S., Aydemir O., Memis A., VARLI S.

ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, cilt.114, 2022 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 114
  • Basım Tarihi: 2022
  • Doi Numarası: 10.1016/j.engappai.2022.105032
  • Dergi Adı: ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, Applied Science & Technology Source, Communication Abstracts, Computer & Applied Sciences, INSPEC, Metadex, Civil Engineering Abstracts
  • Anahtar Kelimeler: Airport turnaround control system, Turnaround action monitoring, Airport vehicle detection, Turnaround action detection, Airport ground services, TRACKING, OPERATIONS, PREDICTION
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

As it is widely known, several ground services are provided by the airports for the domestic and internationalflights of the commercial passenger aircraft. Some of these services are conducted during the period calledas the turnaround which starts with the parking of the aircraft in the aprons before the flight and ends withtheir leave from the aprons for the flight. Turnaround processes achieved in short time periods allow using thelimited airport resources including the service vehicles and staff effectively. In addition, commercial reputationlosses and financial losses that may arise from delays can be reduced as well as the delay-associated turnaroundpenalties. In this article, a deep learning and computer vision based system that detects and allows monitoringthe airport service actions is proposed. The proposed system is capable of analyzing all the primary groundservices for an aircraft parking on its apron by employing the RGB video frame sequences obtained from asingle fixed camera focusing on the apron. In the service detection and analysis modules of the proposed airportground service analysis system, some deep learning-based subsystems and in-house-developed algorithms wereincluded and utilized. For the training of the machine learning models, a study-specific dataset was used andthe constructed learning models were evaluated on real-life cases. Experimental results obtained as a result ofthe performance evaluations show that the proposed system is quite successful with precision rates over 90%in the detection and analysis of the airport ground services. This study is one of the limited research studiesin which deep learning and computer vision techniques have been applied to detect and analyze the groundservice actions. The proposed system is also capable of real-time data processing/analysis and concurrentservice action monitoring. Furthermore, it allows monitoring when the service is received by stamping thetimes of service start/end. In a consideration of industrial relevance or operational perspective, such a systemmay facilitate the airport ground service management noticeably and reduce the delay-associated costs causedby the timing of the ground services