EPILOTS A SYSTEM TO PREDICT HARD LANDING DURING THE APPROACH PHASE OF COMMERCIAL FLIGHTS

EPILOTS A SYSTEM TO PREDICT HARD LANDING DURING THE APPROACH PHASE OF COMMERCIAL FLIGHTS

1 Sindhu S L, 2 Tejaswini S M,

[1] Assistant Professor, Department of MCA, BIET, Davangere

[2] Student, Department of Master of Computer Applications, BIET, Davangere

ABSTRACT

In aviation safety, preventing accidents is paramount, and more than half of all commercial aircraft operation accidents could be avoided through the timely execution of a go- around maneuver. Recognizing this, we have developed a cockpit-deployable machine learning system designed to aid flight crews in making go- around decisions by predicting the likelihood of a hard landing. This system leverages a hybrid approach that incorporates features modeling the temporal dependencies of various aircraft variables, which are then input into a neural network for analysis. Our study utilized an extensive dataset comprising 58,177 commercial flights to train and validate the predictive model. The results demonstrate that our system achieves an average sensitivity of 85% and an average specificity of 74% at the critical go-around decision point. Sensitivity, in this context, refers to the model's ability to correctly identify flights that would result in a hard landing, while specificity indicates the model's accuracy in recognizing flights that would not require a go-around. The significance of these metrics lies in their impact on operational safety. High sensitivity ensures that the system effectively flags potential hard landings, prompting timely go-around decisions that can avert accidents. Meanwhile, adequate specificity minimizes unnecessary go-arounds, thereby maintaining operational efficiency and reducing the risk of other complications. Our approach represents a significant advancement over existing methodologies by integrating real-time data and advanced machine learning techniques. This enables more accurate and reliable recommendations, making it a valuable tool for flight crews. The cockpit-deployable nature of the system ensures that it can be seamlessly integrated into existing flight operations, providing real-time support where it is needed most. In conclusion, our machine learning-based recommendation system for go-around decisions not only enhances flight safety but also optimizes operational efficiency, offering a robust solution to reduce the aviation industry's accident rate.

Keywords: flight safety, cockpit-deployable