Smartphone-based solutions to monitor and reduce fuel consumption and CO2 footprint : final report.

Cetin, Mecit; Ustun, Ilyas; Nadeem, Tamer; Nguyen, Duc; Rakha, Hesham A. · 2016 · ROSA P / TranLIVE. University of Idaho

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 research addresses the need for accurate, large-scale monitoring of fuel consumption (FC) and CO2 emissions in transportation networks, which account for a significant portion of greenhouse gas emissions. The study aims to develop smartphone-based solutions that leverage low-energy sensors—such as accelerometers, gyroscopes, and compasses—rather than relying heavily on GPS, which suffers from high power consumption and poor accuracy in urban environments. The primary objectives were to develop algorithms for detecting travel modes and vehicle operating states, evaluate the effectiveness of emission estimation at various probe vehicle market penetration levels, and create shortest-path algorithms for eco-friendly routing. The methodology involved developing and testing machine learning models using both field data and simulations. For vehicle stop detection, the researchers employed Support Vector Machines (SVMs), Hidden Markov Models (HMMs), and Changepoint Detection Methods (CDMs) using high-resolution accelerometer data. Field data were collected via a custom Android application on smartphones paired with On-Board Diagnostics (OBD) devices, which provided ground-truth speed data. For transportation mode recognition, the team collected 25 hours of data from ten participants across five modes (walking, running, biking, driving, and bus riding) using smartphone sensors. They utilized supervised learning techniques, including Random Forest (RF) and SVM, selecting features via the Minimum Redundancy Maximum Relevance (mRMR) method. Additionally, simulation data were generated to assess estimation errors at different market penetration levels, and shortest-path algorithms were modified to minimize both travel time and fuel consumption. The results demonstrated high accuracy in sensor-based detection. Travel mode detection achieved an average accuracy of approximately 94% across all five modes, with Random Forest and SVM performing best; however, distinguishing between car and bus modes remained challenging, with misclassification rates of 4–6%. Vehicle stop detection using SVM-HMM and CDM models proved effective, accurately identifying stationary and motion states with minimal false alarms. Simulation results indicated that estimation errors for fuel consumption dropped dramatically as probe vehicle market penetration increased from zero to 20%, after which improvements were marginal. The modified shortest-path algorithms successfully identified routes that minimized both travel time and fuel consumption. The significance of this work lies in demonstrating the feasibility of using ubiquitous smartphone sensors to collect vast amounts of dynamic travel data without relying on GPS. The developed algorithms enable accurate estimation of individual carbon footprints and support real-time eco-routing applications. These solutions can be integrated into mobility platforms to provide drivers with feedback on emissions, facilitating more environmentally conscious travel decisions and supporting broader environmental initiatives like Dynamic Low Emissions Zones.

Key finding

Travel mode detection achieved an average accuracy of approximately 94 percent, and fuel consumption estimation errors dropped dramatically as probe vehicle market penetration increased to 20 percent.

Methodology

mixed_methods

Sample size: 10

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.

• exposure measurement