Empirical Data-Driven Approach to Eco-Friendly Deceleration

Driver behavior modification is crucial for reducing CO2 emissions from vehicles. Eco-driving promotes fuel-efficient driving practices and offers a promising solution; this is achieved by suggesting fuel-efficient driving operations to drivers. This study explores the optimal strategies for attaining eco-driving. Specifically, we focus on using a large-scale real driving dataset collected from connected vehicles on public roads to provide insights on the most fuel-efficient driving patterns. To derive the optimal driving behavior, we propose a novel data-driven approach that leverages extensive historical driving data. The key advantage of our approach is its ability to reveal optimal driving behavior and estimate total fuel savings using only real data, without relying on virtual simulations or complex models commonly employed in traditional approaches. We validate the effectiveness of our method by optimizing deceleration behavior before signalized intersections. Through experiments conducted at multiple real intersections, we demonstrate the method's potential to reduce total fuel consumption by over 20%.