Enhancing railway operations: machine learning approaches for stopping distance prediction

Ensuring railway safety systems advance is vital for protecting passengers, pilots, and assets. This study concentrates on creating an automatic braking system for Indian Railways to improve safety by accurately forecasting stopping distances during braking maneuvers. Utilizing the Indian Railways dataset, a thorough analysis of braking data is conducted, emphasizing instances of loco pilot intervention in brake deployment, covering normal braking scenarios and emergencies triggered by the Kavach system. The proposed approach entails filtering the dataset to identify braking instances where loco pilot intervention affects brake application. Subsequently, machine learning regression techniques are utilized to forecast stopping distances based on the filtered brake deployment data. Various regression methods, including linear, polynomial, support vector, and random forest regression, are applied and assessed for their efficacy in accurately predicting stopping distances. The outcomes of this research not only aid in the development of an automatic braking system for Indian Railways but also offer valuable insights into selecting optimal regression methods for forecasting stopping distances in real-world railway operations.