Data Centers on Wheels: Emissions From Computing Onboard Autonomous Vehicles

Abstract

While much attention has been paid to data centers' greenhouse gas emissions, less attention has been paid to autonomous vehicles' (AVs) potential emissions. In this work, we introduce a framework to probabilistically model the emissions from computing onboard a global fleet of AVs and show that the emissions have the potential to make a non-negligible impact on global emissions, comparable to that of all data centers today. Based on current trends, a widespread AV adoption scenario where approximately 95% of all vehicles are autonomous requires computer power to be less than 1.2 kW for emissions from computing on AVs to be less than emissions from all data centers in 2018 in 90% of modeled scenarios. Anticipating a future scenario with high adoption of AVs, business-as-usual decarbonization, and workloads doubling every three years, hardware efficiency must double every 1.1 years for emissions in 2050 to equal 2018 data center emissions. The rate of increase in hardware efficiency needed in many scenarios to contain emissions is faster than the current rate. We discuss several avenues of future research unique to AVs to further analyze and potentially reduce the carbon footprint of AVs.

An attempt to determine the impact of the implementation of autonomous vehicles on a larger scale on the planning of city transport systems

“…The features of the automatic movement of electric self-driving cars can neutralize their environmental advantages. A statistical model developed by MIT researchers shows that soon the energy required for a fleet of electric self-driving cars could produce greenhouse gases equal to those produced by all the data centers in existence today (Sudhakar, Sze, & Karaman, 2023). There are physical places incorporating large computer arrays, and these computers support countless programs worldwide.…”

“…It is estimated that these facilities emit 0.3% of the total carbon emissions, which is approximately equal to the annual carbon emission of Argentina (Sudhakar, Sze, & Karaman, 2023). Researchers estimate that this level of energy is required by self-driving technology for 1 billion vehicles that drive only an hour a day (Sudhakar, Sze, & Karaman, 2023). Comparatively, global road traffic consists of approximately 1.5 billion vehicles (Sudhakar, Sze, & Karaman, 2023).…”

“…Researchers estimate that this level of energy is required by self-driving technology for 1 billion vehicles that drive only an hour a day (Sudhakar, Sze, & Karaman, 2023). Comparatively, global road traffic consists of approximately 1.5 billion vehicles (Sudhakar, Sze, & Karaman, 2023). Electric vehicle computers should consume below 1.2 KW of computational power in 90% of created models to remain within the current data center propagation range; this is something we cannot simply achieve, considering the capabilities of existing hardware (Sudhakar, Sze, & Karaman, 2023).…”

A Full-Stack View of Probabilistic Computing With p-Bits: Devices, Architectures, and Algorithms

“…In the absence of steady transistor scaling and energy improvements, training and maintaining large-scale machine learning models in data centers have become a significant energy concern [1]. The widespread implementation of AI, particularly in industries such as autonomous vehicles [2], is an indication that the energy crisis caused by large-scale machine learning models is not just a data center problem, but a global concern.…”

Peak Age of Information Distribution for Edge Computing With Wireless Links

“…Cooperative driving is one of the major examples of this trend [4]: automated vehicles require huge neural networks to integrate the data from disparate sensors into a coherent picture, and the emissions from on-board computers may become a significant carbon emitter over the next decade [5]. The efficiency gains from deploying Mobile Edge Computing (MEC), and the possibility of coordinating vehicles in an area [6], has led to a significant amount of research on computational offloading [7] in this context.…”