Energy Consumption Optimization for Autonomous Vehicles via Positive Control Input Minimization

TL;DR

This paper presents ECO+, a convex trajectory optimization framework for autonomous vehicles that minimizes positive control input to effectively reduce energy consumption while maintaining safety and comfort.

Contribution

The paper introduces a novel surrogate model based on positive control input and a convex optimization framework for energy-efficient AV trajectory planning.

Findings

ECO+ reduces energy consumption compared to baseline methods.

ECO+ is computationally efficient and scalable.

Initializing nonlinear solvers with ECO+ offers minimal additional benefits.

Abstract

Autonomous vehicles (AVs) present a unique opportunity to improve the sustainability of transportation systems by adopting eco-driving strategies that reduce energy consumption and emissions. This paper introduces a novel surrogate model for energy and fuel consumption that minimizes Positive Control Input (PCI). Unlike conventional objectives such as squared acceleration, which often misrepresent actual energy usage, PCI provides a more accurate and optimization-friendly alternative. Building on PCI, we propose ECO+, a convex, time-based trajectory optimization framework that ensures safety and passenger comfort while optimizing energy use for AVs approaching an intersection. To improve computational efficiency, quadratic resistive forces are approximated using piecewise affine segments, resulting in a linear programming formulation. ECO+ is validated using empirical fuel and electric energy models and benchmarked against established optimization strategies, including a state-of-the-art nonlinear solver. Simulation results demonstrate that ECO+ consistently outperforms baseline methods in reducing energy consumption, even under strict comfort constraints and in scenarios involving a leading vehicle. Moreover, initializing a nonlinear solver with ECO+ yields only marginal gains, indicating that ECO+ is effective as a standalone eco-driving strategy. These findings highlight ECO+ as a practical, scalable, and computationally efficient solution for enhancing the sustainability of autonomous urban mobility systems.