A Dynamic Model for Bus Arrival Time Estimation based on Spatial Patterns using Machine Learning

The notion of smart cities is being adapted globally to provide a better quality living. A city's mobility component focuses on providing smooth and safe commuting for its residents promotes eco-friendly sustainable alternatives such as public transit (bus). Among several applications, system that provides up-to-the-minute information like bus arrival, travel duration, schedule, etc., improves the reliability services. Still, this application needs live traffic flow, accidents, events, location buses. Most lack infrastructure these data. In context, arrival prediction model proposed forecasting time using limited data sets. buses spatial characteristics are used study. One routes Tumakuru city service, Tumakuru, India, selected divided into two patterns: sections with intersections without intersections. machine learning XGBoost modeled both patterns individually. dynamically predict developed preceding trip estimate at downstream stop. performance models compared based R-squared values predictions made, established superior results. It suggested in study area. can also be extended other similar traffic-related infrastructure.