Can driving styles prove the smarter route to better fuel economy and emissions?

The relevance of official fuel economy figures is still a perennial topic across the automotive industry, and one of the hardest to answer directly takes into account the human factor. Some claim that the official figures are in fact perfectly valid, if only the average driver wasn’t so lead-footed. But is there any truth in this train of thought?

Thinking laterally, one way to assess this question is to consider a scenario where the UK parc features a high penetration of connected and autonomous vehicles (CAVs).  By taking the driver out of the loop an autonomous vehicle could deliver a more “responsible” and consistent driving style.  This would in turn consume less fuel and so deliver significant reductions in nitrogen oxide (NOx) and carbon dioxide (CO2) emissions, then perhaps the driving style hypothesis is true.

To look at this very question Emissions Analytics and Imperial College London collaborated on the “Optimised Vehicle Autonomy for Ride and Emissions” feasibility project, supported by the Centre for Connected and Autonomous Vehicles and Innovate UK.

The research methodology started with Emissions Analytics performing on-road PEMS tests of 21 vehicles to its standard EQUA Index protocol across the UK and Germany.  This provided second-by-second driving characteristics and the simultaneous emissions, allowing an emissions map for each car to be created.

These were then integrated with VISSIM ( traffic simulation software to model the effects of traffic flow and driving behaviour on the emissions. The simulation covered three main factors: vehicle driving strategy, communication between vehicles and infrastructure, and the penetration of CAVs. This included scenarios were a CAV followed a normal vehicle, a non-CAV followed a CAV, and CAV following another CAV.

We also factored in the ability for CAVs to communicate with traffic lights allowing them to approach a junction knowing when they were going to change, and so avoid coming to a complete stop.

In essence, the virtual CAVs were configured to mimic a more cautious and gentle
driver with better anticipation of the road ahead and smoother style of driving.  The benefit of the approach was we could model a world of 100% CAV penetration, but also the transition to that point via a mixed fleet of CAVs and non-CAVs.

The headline results are that, with 100% CAV penetration with congested traffic, NOX and CO2 could be reduced by approximately 20%. The preliminary analysis assumes a simplified scenario with only diesel cars, so the results are not yet reflective of the current UK fleet. It is plausible that petrol vehicles would show a similar CO2 reduction, while NOX emissions are typically low NOX whatever. Overall, the results show that smoother traffic flow could have significant benefits in urban areas.

This suggests that optimised driving styles can deliver lower vehicle emissions.  However, according to our EQUA Index (, on average diesel NOx emissions are approximately five times the regulated limit on average (399 mg/km), and CO2 emissions are 40% above official values (based on the New European Driving Cycle).  This means that even if driven more responsibly the average vehicle will only reduce its NOx exceedances to about a factor of four, and the CO2 exceedances to 30%.  While this is an improvement, and underlines the benefits of driver training initiatives, it strongly suggests that the driver is not responsible for the majority of the emissions exceedances observed.

The chart below shows the reduction in NOx, as the penetration of CAVs increases from 0% to 100%. The benefits are seen is a broadly linear way as penetration increases, although the majority of NOx emissions still come from non-CAV vehicles even when CAV penetration passes 60%.

In summary, even if the driver and all associated “bad” driving habits are eliminated, the current fleet of vehicles will still well exceed official CO2 values, and diesel vehicles will exceed the NOX limits.  Modifying driver behaviour can offer a valuable mitigation to these exceedances, but real-world emissions are still primarily determined by vehicle selection.


Emissions Analytics