Engineering Informatics for Aviation Safety: Machine Learning-Based Prediction of Bird Strikes Using a Model-Based CPSS Design

Abstract

Bird strikes on aircraft present a growing concern to wildlife preservation efforts and pose safety and economic challenges in aviation and urban design that amalgamate a sociotechnical system. Recent evidence suggests that artificial light at night is a contributing factor to bird strikes, especially during migration seasons, when birds are more likely to be disoriented by bright urban lighting. In this article, we explore the interplay of factors influencing bird strikes and present a machine learning-based predictive modeling approach informed by cyber-physical-social systems (CPSS). We emphasize the CPSS paradigm as a lens for modeling a range of systems engineering problems where evolving geospatial, temporal, and societal constraints characterize the environment. We underscore the interactions between the cyber, physical, and social spaces in model development and feature selection, providing a model-based system design with clear interpretations for public policy design and decision-making. Through exploratory data analysis and predictive modeling, we identify significant patterns and trends in bird strikes and develop a model to predict bird strikes at a given time and location, with implications for urban planning, lighting design, and aviation safety. We emphasize the potential of these predictive models to inform decisions in public policy and aerospace safety, aiming to mitigate bird strikes while accounting for ecological and industrial factors through model-based systems design. We envision this providing an actionable step toward designing evolving cyber-physical-social systems and engineering informatics for public policy.