Optimizing Traffic Controllers along the MLK Smart Corridor Using Reinforcement Learning and Digital Twin

Abstract

With advancements in Intelligent Transportation Systems (ITS), sensors, and computing resources, several cities across the world are investing in the development of smart/connected corridors. These corridors are being equipped with advanced sensors that provide real-time, high-resolution data from the corridor and enable vehicle-to-infrastructure (V2I) and vehicle-to-vehicle (V2V) communications. The objective of this study is to optimize signal timings for one such smart corridor – MLK Smart Corridor – in Chattanooga, Tennessee, USA with respect to fuel and energy consumption (represented by Fuel Consumption Intersection Control Performance Index, EcoPI, that determines the excess fuel consumption due to stops and delays caused by traffic controllers).