How EcoTrace Calculates CO2e for HGV Fleets

The limitation of average emission factors

A standard emission factor for an HGV diesel vehicle is an industry average. It does not account for the load your vehicle was carrying, the gradient of the route, the efficiency characteristics of that specific engine, or the driving conditions on that journey.

For logistics operators providing carbon data to enterprise customers — or reporting their own Scope 3 emissions under CSRD — the gap between an estimated average and a journey-specific calculation is meaningful. ISO 14083 distinguishes between primary activity data (calculated from actual operational inputs) and secondary data (derived from population averages). The data quality tier determines the credibility of the disclosure.

Physics-based CO2e calculation — what it means in practice

EcoTrace is developing the Scientific Carbon Validation Engine (SCVE), a software system that calculates CO2e at the level of individual HGV journeys.

The SCVE uses Physics-Informed Neural Networks (PINNs) — a class of machine learning architecture that embeds physical governing equations directly into the training process. For a heavy goods vehicle, these equations describe how traction force, aerodynamic drag, road gradient, rolling resistance, and engine efficiency interact to determine fuel consumption at every point along a journey.

From this, the system derives:

• Fuel consumption per trip — calculated from vehicle dynamics, not estimated from distance
• CO2e per trip — derived from calculated fuel consumption and established emission factors
• Vehicle-specific parameter calibration — engine efficiency and rolling resistance coefficients calibrated to each vehicle from its own telemetry data, without requiring manual measurement

The inputs the system processes are those most modern HGV fleets already collect: GPS speed data, route elevation, and fuel records.

Current development stage

EcoTrace is at the R&D stage. Proof-of-concept experiments conducted in early 2026 on synthetic datasets demonstrated that the core PINN architecture achieves sub-1% Mean Absolute Percentage Error (MAPE) for fuel consumption prediction on variable-gradient routes — within the accuracy target specified for the system.

The full system — incorporating real vehicle telemetry from operational fleets, multi-vehicle calibration, and alignment with ISO 14083 — is the subject of ongoing R&D development.

We are not yet offering a commercial product. We are building the technical foundation for one.

Designed for logistics operators and sustainability consultancies

The SCVE is being developed as a B2B software API — accessible to logistics operators directly, and to sustainability consultancies and ERP implementation partners managing carbon reporting for logistics clients.

Target users include:

• Sustainability managers at logistics and freight companies who need journey-level CO2e data for customer reporting and internal reduction tracking
• Sustainability consultancies providing ISO 14083 and CSRD advisory services to logistics clients
• ERP and TMS implementation partners integrating carbon data into existing logistics management infrastructure

Follow our R&D progress

EcoTrace is in active R&D. If you are a logistics operator, sustainability consultancy, or technical partner interested in following our development — or in participating in our planned operational pilot — we would welcome the conversation.

EcoTrace Green Technologies Ltd — Company No: 17180344 — London, United Kingdom. The SCVE is under active R&D development. No commercial product is currently available. Results cited refer to proof-of-concept experiments conducted on synthetic datasets.