Artificial Intelligence Enabled Model for Real Time Prediction of Aircraft Touch-Down Point
2022-26-1192
05/26/2022
- Event
- Content
- Approach and Landing are critical phases of flight. On a typical flight with a duration of 1.5 hours, the landing phase accounts for approximately 1% of total time. Statistics and past data indicate that landing accidents, in jet powered aircraft, result in about 20% of aviation related accidents / incidents. Criticality of landing phase is further accentuated when an aircraft lands on table top runways with short runway during adverse climatic conditions. Studies have indicated that risk of runway overruns increase by 55% due to ‘long landings’. The proposed idea is an attempt to reduce long landings by accurately predicting the touchdown point and generating an alert to ATC in real time, in case of deviations, thereby, avoiding runway excursions caused by long landings. The proposed solution makes use of ADS-B out data received from aircraft with additional parameters gathered from a multitude of sensors in real time and predicts the precise touch down point based on trajectory prediction and historical data analysis from past landings at the airport. The proposed methodology will help reduce runway excursion incidents attributable to long landings resulting in enhanced flight safety. The proposed approach also predicts the trajectory onboard using flight computers combined with data component from airport to provide an accurate prediction of the touch down point well before Minimum Descent Altitude is reached, thereby enabling pilots to take a go-around decision, in case probability of long landing is above a certain threshold. The paper highlights and explores various machine learning models for suitability to get the optimal glide path and touchdown point prediction.
- Citation
- Padmanabhan, S., and Sunkara, B., "Artificial Intelligence Enabled Model for Real Time Prediction of Aircraft Touch-Down Point," SAE Technical Paper 2022-26-1192, 2022, .