How Important is it for Self-Driving Cars to be Electric?

May 3, 2018 | 12:04 pm
A Waymo self-driving car on the road in Mountain View, CA, making a left turn. CC-BY-2.0 (Wikicommons).
David Reichmuth
Senior Engineer, Clean Transportation Program

The rapid development of self-driving technology has raised many important questions such as the safety of automated vehicles (AVs) and how they could radically alter transportation systems. These are critical questions, but AVs also have the potential to result in significant changes to the global warming emissions from personal transportation.

An interesting recent study from the University of Michigan and Ford Motor Company lays out the details of the likely changes in emissions from using an AV system on both electric and gasoline cars. The main takeaway from the study is that adding AV equipment to a car adds weight, aerodynamic drag, and electrical power consumption that leads to increased fuel consumption. There is the potential to offset emissions from more efficient driving by connected and automated vehicles, but by far the largest impact on emissions is the choice of fuel: gasoline versus electricity.

Direct emissions versus behavioral and usage changes

Switching from human control to fully automated driving will have direct effects on emissions as well as changes to the amount we use vehicles. Direct emissions changes include reductions in efficiency from factors like increased drag from sensor equipment and the power consumption of required computing and communications equipment. Positive direct impacts could include more efficient driving, such as smooth and precise acceleration control in an automated system.

Automation will also change how we use cars and how much we use them, indirectly affecting emissions, though the effect of AVs on these indirect emissions is much more speculative. While some changes, like “right-sizing’ (for example, having smaller one or two occupant cars available for solo trips), could decrease emissions, many of the usage changes considered would increase vehicle usage and therefore emissions. Making long distance driving easier or more productive could encourage people to live farther from their jobs. Having fully automated vehicles will mean more people can use a car. The elderly, blind, youth, and people with disabilities could switch from transit to a car, or simply add trips that would not have been able to happen otherwise. While many of these uses of AVs would be beneficial, it’s important to understand the potential emissions from AVs and how we could minimize the total contribution of global warming pollution from personal transportation.

That’s why this new study is important: it lets us at least estimate the direct, short-term implications of AV technologies on emissions. While it doesn’t examine the potential impacts of driving more, it does shed light on the direct effects of adding these new features to cars.

AV equipment increases fuel consumption, especially for gasoline vehicles

Focusing on the physical changes to the vehicle, the addition of self-driving and sensor equipment has three major changes to the fuel consumption (and therefore emissions) of the AV. First, the additional weight of the equipment decreases efficiency. Second, AVs that have sensor equipment like cameras and LiDAR (laser-based imaging) often require side bulges and roof-mounted equipment pods. Like a conventional cargo rack, these additions are detrimental to fuel economy as they increase the vehicle’s aerodynamic drag. Lastly, the sensors and computing equipment that enable self-driving require additional electrical power beyond a conventional vehicle. For a gasoline car, this means added load on the engine to power an alternator (and therefore higher gasoline consumption), while a battery electric car will have reduced overall driving efficiency (and therefore shorter range between recharges).

Waymo’s AV minivan adds sensors and computing systems that increase weight, drag, and electrical power consumption. This model was used as an example of a ‘large’ sized AV system in the referenced study. Image source: Waymo

The researchers from Michigan and Ford examined three sizes of AV systems that could be added to vehicles: an AV system with sensors like a Tesla Model S, a medium-sized system with smaller external sensors similar to a Ford AV prototype, and finally a large AV system modeled after Waymo’s modified Chrysler Pacifica AV. While all AV systems have a negative impact on fuel consumption and emissions, the largest impact is seen in the increased drag from the large AV system.

AV systems can increase global warming emission attributed to driving. The largest impact is seen on larger AV systems due to drag from the sensor units. Adapted from Gawron et al

Improved driving behavior and other savings from AVs are possible in the long run

The study also points out the possibility of fuel savings from having self-driving and connected cars. These savings could come from several sources. For example, AVs could have more efficient acceleration and braking (“eco-driving”), especially if they are communicating with other cars to anticipate speed changes in traffic. AVs could also communicate with infrastructure like traffic signals to reduce idling and stop-and-go driving. On highways, groups of connected AVs could drive much closer together than a human driver could. This ‘platooning’ technology can increase fuel efficiency by reducing aerodynamic resistance, similar to the drafting that competitive cyclists and NASCAR drivers use to save energy. There is also a potential for AV technology to increase fuel consumption because cars could potentially drive safely on the highway at higher speeds and high speeds reduce efficiency.

These factors are currently harder to quantify than the impact of the AV equipment, and some of the potential benefits require having most or all cars on the road be at least connected, if not fully automated. For example, platooning would require multiple AVs traveling on the same roadway at the same time, which would require a critical mass of AVs to be deployed. The researchers in this study estimate a potential emissions savings on average of 14 percent from these technologies if fully implemented. However, they do not consider changes to vehicles that are already producing some of these benefits, such as improved aerodynamics (which gives some of the same benefits as platooning) or stop-start systems (which already act to reduce some of the adverse impacts of stop-and-go traffic and intersections).

Early AV models are more likely to have higher emissions

The study also considered the impact of the much more power-hungry equipment used in early developmental AV systems. For example, early prototypes have been reported to require in excess of 2,000 W of power, mostly for on-board computing. Increased computer power requirements in these early prototypes, for example going from the from these early AVs (see table). This is especially true for the less-efficient gasoline-engine driven vehicles, where increased electric power requirements would increase emissions over 60 grams CO2 equivalent per mile.  That’s equal to reducing the fuel economy of a 35MPG car to 29MPG, or like adding the emissions from running 10-25 iMac computers using a gasoline generator for every car. Since early AVs will not have enough numbers on the road to take advantage of platooning and connected vehicle savings, it is very likely that in the near-term AVs will contribute higher net emissions than a conventionally driven vehicle using the same fuel.


Emissions from AV system’s electricity use. Baseline system is 200W computer system, prototype uses 2,000W computing system.

AV system size

Baseline AV system,

battery electric vehicle (gCO2eq/mi)

Baseline AV system, gasoline vehicle (gCO2eq/mi)

Prototype AV system, battery electric vehicle (gCO2eq/mi)

Prototype AV system, gasoline vehicle (gCO2eq/mi)



Switching from gasoline to electricity is by far the most important factor in reducing emissions


The choice of fuel (gasoline versus electricity) is the most important choice for reducing emissions. Emissions estimates based on Ford Focus gasoline and battery-electric models and includes ‘well-to-wheel’ emissions for fuel production, distribution, and use in the vehicle. Emissions related to vehicle or AV system production are not included in this chart.

The most important determinant of direct emissions from vehicles is not the AV system, but is the choice of gasoline or electricity. Choosing a electric vehicle instead of the gasoline version for this analysis reduces global warming emissions from 20 to over 80 percent, depending on the emissions from electricity generation. The addition of AV equipment only increases this difference, making it clear that electric drive is required to have AVs that maximize emissions reductions.

What will the future hold? Some AV companies, like Waymo (spun off from Google) and Cruise Automation (partnered with General Motors) are using EVs and have plans to continue using electric drive in their AVs. Other companies have been less progressive, such as Ford announcing that they anticipate using gasoline-only hybrids for their AVs. If AVs have the transformative effect on mobility and safety that many predict, it will be vital to encourage the use of cleaner electricity instead of gasoline in these future vehicles.