A Decentralized Time- and Energy-Optimal Control Framework for Connected Automated Vehicles: From Simulation to Field Test

TL;DR

This paper presents a comprehensive implementation and validation of a decentralized, optimal control framework for connected automated vehicles, demonstrating improved energy efficiency and travel time through simulation, testing, and real-world field trials.

Contribution

The paper develops and validates a novel decentralized control framework for CAVs, integrating simulation, hardware-in-the-loop, virtual reality bench-testing, and field testing in a real vehicle.

Findings

Significant energy savings and reduced travel time in tests.

Successful integration of control framework across multiple testing stages.

Effective real-world deployment in a field test environment.

Abstract

The implementation of connected and automated vehicle (CAV) technologies enables a novel computational framework for real-time control aimed at optimizing energy consumption with associated benefits. In this paper, we implement an optimal control framework, developed previously, in an Audi A3 etron plug-in hybrid electric vehicle, and demonstrate that we can improve the vehicle's efficiency and travel time in a corridor including an on-ramp merging, a speed reduction zone, and a roundabout. Our exposition includes the development, integration, implementation and validation of the proposed framework in (1) simulation, (2) hardware-in-the-loop (HIL) testing, (3) connectivity enabled virtual reality based bench-test, and (4) field test in Mcity. We show that by adopting such inexpensive, yet effective process, we can efficiently integrate and test the controller framework, ensure proper connectivity and data transmission between different modules of the system, and reduce uncertainty. We evaluate the performance and effectiveness of the control framework and observe significant improvement in terms of energy and travel time compared to the baseline scenario.