Estimating Light-Duty Vehicle Fuel Consumption and CO2 Emissions via OBD-II Speed-Density Modeling: A Field Demonstration

Laboratory-based certification cycles systematically underestimate real-world fuel consumption and CO2 emissions. On-board diagnostics (OBD-II) telemetry offers a low-cost alternative, yet most published approaches rely on mass air flow (MAF) sensors absent from many modern vehicles. This study validates a speed-density air-mass estimation method on a naturally aspirated RON 95 gasoline passenger car (1368 cm3, Euro 6) across seven drive cycles recorded over three measurement days in northwestern Türkiye, covering 609.6 km of highway, urban, and mixed conditions. Instantaneous air mass flow was estimated from four standard OBD-II PIDs—manifold absolute pressure, engine speed, intake air temperature, and fuel trim corrections—using the ideal gas law applied to actual engine displacement. Results were validated against pump-measured fill-up volumes. The speed-density model achieved errors of −3.6% to +4.3% across individual segments (combined error: −0.5%), outperforming the vehicle’s onboard trip computer, which exhibited errors of −10.6% to +14.6%. Derived CO2 intensities ranged from 125.0 to 166.4 g/km, with a combined average of 147.2 g/km (pump reference: 147.9 g/km). Urban driving produced approximately 15% higher specific emissions than highway driving. These results demonstrate that a physics-based speed-density model can achieve within ±5% trip-level accuracy across diverse real-world conditions without machine learning, bespoke calibration, or a physical MAF sensor.