Drone, self-driving vehicle energy savings overstated, reports find

Autonomous vehicles and drones could reduce greenhouse gases and energy use in the transport sector, two reports show, but the savings are in many cases modest.

A modelling study published in the journal Nature Communications found that delivery of packages weighing up to five kilograms by drones, such as quadcopters, required significantly less energy than delivery by small truck. 

However, the limited flight range of the drones – currently around four kilometres – means that any drone-based delivery system will require a network of additional warehouses, the construction and operation of which will add to energy use and gas emission totals.

The analysis, conducted by a team led by Joshuah Stolaroff of the Lawrence Livermore National Laboratory in California, US, found that a delivery system based on drones – a model favoured by Amazon and predicted to become commonplace in the next few years – did not always perform better than one using ground-based transport.

“There are plausible scenarios where drones lead to overall higher energy use and greenhouse gas emissions compared to ground vehicles,” they write.

One of these was the increased energy expenditure required to build and maintain warehouses. Another was continued improvement in the energy efficiency of trucks, and a third (counter-intuitively) was increased efficiency of the drones themselves allowing them to carry heavier packages for longer.{%recommended 1047%}

Overall, Stolaroff and his colleagues conclude, positive outcomes from drone-based systems will rely heavily on two underpinning factors – a decrease in the carbon intensity of metropolitan electricity systems, and increased energy efficiencies in the operation of commercial buildings.

Meanwhile, a second report, published in the journal Environmental Science and Technology, finds that autonomous ground-based vehicles – self-driving cars – deliver energy use and greenhouse gas emission savings because they operate much more efficiently than vehicles controlled by humans.

However, those savings are significantly offset by the extra weight and aerodynamic drag produced by the inclusion of the cameras, computers, navigational equipment and support structures needed to achieve autonomy in the first place.

The study, led by Gregory Keoleian of the University of Michigan in the US, also found that wireless data transmission for autonomous vehicle map systems required hefty additional energy consumption.

Previous modelling by the US National Renewable Energy Laboratory concluded that autonomous vehicles should be 14% more efficient than human-driven equivalents.

Keoleian’s team found that a more realistic real-world estimate was 9%.