An adaptive big data weather system for surface transportation

Abstract

Operating modern multi-modal surface transportation systems are becoming increasingly automated and driven by decision support systems. One aspect necessary for successful, safe, reliable, and efficient operation of any transportation network is real-time and forecasted weather and pavement condition information. Providing such information requires an adaptive system capable of blending large amounts of observational and model data that arrives quickly, in disparate formats and times, and blends and optimizes their use via expert systems and machine-learning algorithms. Quality control of the data is also essential, and historical data is required to both develop expert-based empirical algorithms and train machine learning models. This paper reports on the open-source Pikalert® system that brings together weather information and real-time data from connected vehicles to provide crucial information to enhance the safety and efficiency of surface transportation systems. This robust framework can be applied to a diverse array of user community specifications and is designed to rapidly ingest more, unique data sets as they become available. Ultimately, the developmental framework of this system will provide critical environmental information necessary to promote the development, growth, refinement, and expanded adoption of automated and connected multi-modal vehicular systems globally.