Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Traffic-Aware Eco-Driving Control in CAVs via Learning-based Terminal Cost Model

Mehmet Fatih Ozkan, Dennis Kibalama, Jacob Paugh, Marcello Canova, Stephanie Stockar

TL;DR

The paper addresses energy efficiency in eco-driving for CAVs by integrating macroscopic traffic effects beyond the optimization horizon into a neural-network-based terminal cost within a model predictive control framework. It introduces two ensembles—T-Ag-NN and T-Aw-NN—to generate terminal costs, with T-Aw-NN incorporating traffic jams via Greenshields-based speed limits and lead-vehicle projections. Using full-route DP data to train the networks and Rollout-based MPC, the Ensemble-NN MPC demonstrates smoother trajectories and up to 6.4%–6.5% energy savings over traffic-agnostic approaches, at the cost of modest travel-time changes, highlighting practical potential for real-time traffic-aware eco-driving in urban CAVs. The approach bridges microscopic control with macroscopic traffic dynamics, enabling more robust performance under dynamic traffic conditions and jams.

Abstract

Connected and Automated Vehicles (CAVs) offer significant potential for improving energy efficiency and lowering vehicle emissions through eco-driving technologies. Control algorithms in CAVs leverage look-ahead route information and Vehicle-to-Everything (V2X) communication to optimize vehicle performance. However, existing eco-driving strategies often neglect macroscopic traffic effects, such as upstream traffic jams, that occur outside the optimization horizon but significantly impact vehicle energy efficiency. This work presents a novel Neural Network (NN)-based methodology to approximate the terminal cost within a model predictive control (MPC) problem framework, explicitly incorporating upstream traffic dynamics. By incorporating traffic jams into the optimization process, the proposed traffic-aware approach yields more energy-efficient speed trajectories compared to traffic-agnostic methods, with minimal impact on travel time. The framework is scalable for real-time implementation while effectively addressing uncertainties from dynamic traffic conditions and macroscopic traffic events.

Traffic-Aware Eco-Driving Control in CAVs via Learning-based Terminal Cost Model

TL;DR

The paper addresses energy efficiency in eco-driving for CAVs by integrating macroscopic traffic effects beyond the optimization horizon into a neural-network-based terminal cost within a model predictive control framework. It introduces two ensembles—T-Ag-NN and T-Aw-NN—to generate terminal costs, with T-Aw-NN incorporating traffic jams via Greenshields-based speed limits and lead-vehicle projections. Using full-route DP data to train the networks and Rollout-based MPC, the Ensemble-NN MPC demonstrates smoother trajectories and up to 6.4%–6.5% energy savings over traffic-agnostic approaches, at the cost of modest travel-time changes, highlighting practical potential for real-time traffic-aware eco-driving in urban CAVs. The approach bridges microscopic control with macroscopic traffic dynamics, enabling more robust performance under dynamic traffic conditions and jams.

Abstract

Connected and Automated Vehicles (CAVs) offer significant potential for improving energy efficiency and lowering vehicle emissions through eco-driving technologies. Control algorithms in CAVs leverage look-ahead route information and Vehicle-to-Everything (V2X) communication to optimize vehicle performance. However, existing eco-driving strategies often neglect macroscopic traffic effects, such as upstream traffic jams, that occur outside the optimization horizon but significantly impact vehicle energy efficiency. This work presents a novel Neural Network (NN)-based methodology to approximate the terminal cost within a model predictive control (MPC) problem framework, explicitly incorporating upstream traffic dynamics. By incorporating traffic jams into the optimization process, the proposed traffic-aware approach yields more energy-efficient speed trajectories compared to traffic-agnostic methods, with minimal impact on travel time. The framework is scalable for real-time implementation while effectively addressing uncertainties from dynamic traffic conditions and macroscopic traffic events.

Paper Structure

This paper contains 11 sections, 9 equations, 4 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Vehicle Dynamics & Powertrain Model
  3. Problem Formulation
  4. Full Route Optimization
  5. Receding Horizon Optimization
  6. Traffic Agnostic Neural Network (T-Ag-NN)
  7. Traffic-Aware Neural Network (T-Aw-NN)
  8. Results and Discussion
  9. Simulation Setup
  10. Simulation Results
  11. Conclusion

Figures (4)

  • Figure 1: Ensemble-NN (T-Ag-NN & T-Aw-NN).
  • Figure 2: Interaction of ego vehicle, traffic proxy and upcoming traffic jam.
  • Figure 3: Comparison of wait-and-see DP, Ensemble-NN and T-Ag-NN MPC solutions in Route 1.
  • Figure 4: Comparison of wait-and-see DP, Ensemble-NN and T-Ag-NN MPC solutions in Route 2.