Real-Time Implementation Comparison of Urban Eco-Driving Controls

Rabinowitz, Aaron I; Ang, Chon Chia; Mahmoud, Yara Hazem; Araghi, Farhang Motallebi; Meyer, Richard T; Kolmanovsky, Ilya; Asher, Zachary D.; Bradley, Thomas · 2024 · ROSA P / Mountain-Plains Consortium

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

This study addresses the lack of comprehensive comparative analysis in the literature regarding autonomous eco-driving controls for connected autonomous vehicles (CAVs). Motivated by the need to reduce greenhouse gas emissions and improve energy economy (EE), the authors investigate how different control strategies perform in real-time urban environments. While manual eco-driving is effective, autonomous implementation offers greater precision, scalability, and the ability to utilize beyond-line-of-sight traffic data. The research aims to evaluate representative control methods—categorized as rule-based, discretized control optimization, and polynomial trajectory optimization—to determine the best trade-off between energy efficiency gains and computational feasibility. The methodology involves a literature review and a simulation-based comparative study using real-world traffic signal data collected from a four-mile route in Fort Collins, Colorado. The authors developed a system framework comprising perception, planning, and plant subsystems. The perception subsystem generates path constraints based on vehicle-to-infrastructure (V2I) signal phase and timing (SPaT) data and vehicle-to-vehicle (V2V) information, creating a "corridor" of allowable positions and speeds. The planning subsystem implements various solvers, including dynamic programming (DP), genetic algorithms (GA), and particle swarm optimization (PSO), within a validated Future Automotive Systems Technology Simulator (FASTSim) model of a Kia Soul EV. The simulations compare these methods against a baseline intelligent driver model (IDM) using different cost functions, including pure acceleration minimization and velocity-sensitive metrics. The results demonstrate that autonomous eco-driving can achieve energy economy improvements ranging from 5% to 15% depending on the method and cost function employed. Dynamic programming (DP) methods yielded the highest energy efficiency improvements but were significantly more computationally expensive, suffering from the "curse of dimensionality" that limits their real-time applicability. In contrast, genetic algorithm (GA) methods presented the most potential for practical real-time implementation, offering a favorable balance between energy savings and run-time efficiency. Particle swarm optimization (PSO) executed faster than DP but underperformed in terms of energy efficiency. Furthermore, the study found that using velocity-sensitive cost functions allowed all tested methods to outperform those relying solely on acceleration minimization. The significance of this work lies in its identification of genetic algorithms combined with road power cost functions as the optimal strategy for generating eco-driving traces for urban battery electric vehicles. This recommendation provides a viable path for real-time implementation that balances computational load with substantial energy savings. The findings suggest that widespread adoption of such autonomous eco-driving controls could have a significant national impact on energy conservation and emissions reduction. The study also highlights the importance of incorporating realistic, time-varying constraints derived from real-world infrastructure data to ensure the practicality of proposed control strategies.

Key finding

Genetic algorithms offer the most promising balance for real-time urban eco-driving by providing significant energy economy improvements with lower computational costs than dynamic programming methods.

Methodology

simulator

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via bulk_ingest_rosap on 2026-05-23 (6 acquisition events logged).

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• eco driving

Information type

What kind of knowledge this paper contributes, grouped by family — independent of topic (what it is about) and method (how it was studied).

• Theoretical Contribution: computational model