Machine Learning-Based Approach for Predicting Hard Landings in Commercial Aircraft

The landing phase, one of the flight phases, is considered the most critical phase due to its significantly higher accident rate compared to other flight phases. A large portion of accidents occurring during the landing phase consists of hard landings. A hard landing is a landing incident defined as the main landing gear impacting the ground with a greater vertical speed and force than a normal landing. The severity of hard landings can vary from minor passenger discomfort to serious aircraft damage, structural failure, or even loss of life. In this study, the decision-making process regarding go-around maneuvers based on hard landing prediction is addressed using Machine Learning methods, specifically Logistic Regression and Random Forest models. Modeling was conducted using a dataset composed of real-time flight parameters, aiming to prevent hard landing incidents and even landing accidents. The calculation results indicate that the developed models provide accurate predictions for hard landing events.