NREL and Volvo scope scale of autonomous driving fuel savings

The US Department of Energy’s National Renewable Energy Laboratory (NREL) and Volvo Cars have partnered in a research project, looking to find out the fuel savings that could be realized using self-driving technologies. As full electrification of new cars is going to take some time, the semi-autonomous systems like advanced cruise control (ACC) are going to appear in internal combustion engine (ICE) vehicles for quite some time, and so NREL and Volvo want to gauge how beneficial these systems might be.

NREL wanted to develop an objective measurement, and so it pulled data from Volvo cars near Gothenburg, on the ‘Drive Me’ project road network, comparing the fuel usage of those using ACC to this not using the technology. The pair did not want to conduct testing in laboratories, as the results would not reflect real-world conditions, and used the data from 18,500 trips taken by Volvo employees and their family members in 93 cars.

The results found a 5-7% saving in fuel, which is a big deal if ACC becomes a standard feature for new cars moving forward. However, due to the low current penetration of ACC, the technology isn’t going to be able to drum up much of a fanfare about its eco-credentials.

Other considerations that the research team want readers to be aware about include the impact of ACC-enabled cars have on other vehicles around them. As these ACC vehicles are typically operating around non-ACC vehicles, simulations show that lots of ACC-enabled cars operating in proximity can worsen traffic due to their lags in speed detection.

This could mean that while one ACC-enabled car saves 5-7% of its fuel consumption, its impact on the rest of the traffic system could offset these improvements – as all the nearby ACC cars are going to be estimating the correct speed themselves, which will slow down the flow.

That’s a study that is going to require a much greater effort on the part of NREL and Volvo, but the problem itself can be solved by using Cooperative-ACC (CACC), which relies on V2X communications to get the ACC cars in sync. In a Safe Road Trains for the Environment (SARTRE) project, which this study points to, CACC could enable a 20% reduction in fuel usage, but this is for vehicles traveling in convoys, rather than general usage driving.

The other consideration that the researchers warn about is the possibility that greater ACC adoption translates into greater use of the vehicles for long trips, where ACC ends up being responsible for people traveling further on the roads than they did before, thanks to its comfort.

The pair are hopeful that more studies can be performed, thanks in part to the greater penetration of vehicle telematics and sensor data collection. The data that semi-automated vehicles collect can be used by researchers to study other traffic conditions or behaviors. This should let them model ‘what if’ scenarios, as well as study how the fuel economy is changed by new policies and penetration rates.

The data should also enable the automakers to start bragging about their technologies in their marketing materials, as well as cozying up to environmental regulators – however, their feet need to be held to the fire if they are cherry-picking data points.

Gothenburg’s Drive Me network spans a 40km route, consisting of highways with no pedestrian or bicycle access. The speed limit is 80kph (50mph), and the route has 130 fixed traffic detector locations, which can track traffic speed and road conditions. TomTom and the Swedish Transport Administration (Trafikverket) were used to provide geographic information about the route, including the gradients of the roads.

The paper defines the mathematical framework that was developed to determine the impact of ACC, which is a little too in-depth for this article. It concludes that it now has a fuel impact analysis framework that can be applied to ACC evaluations. It noted that it found it had insufficient data in some driving conditions, and again stressed that this framework is not going to work in areas where there is high penetration of ACC-enabled cars on the road.

To this end, NREL talks about future work that would include improvements to the traffic flow estimation model used by the framework, and the incorporation of hybrid and electric drivetrains into the data set – to determine the impact of ACC on those vehicles’ energy usage.