Menu

Is 5G Fast Enough for the Edge?

April 26, 2022 1 Comment

Is 5G Fast Enough for the Edge?

By Martin Stiborski, Managing Director, BRESSNER Technology

5G networks are the key for a worldwide digital transformation, impacting AI and edge computing as well as smart cities, while also providing high-speed data transfer in real-time for enterprises across every vertical market segment. New applications are constantly emerging on the market, taking advantage of the cellular networks’ super-low latency and much higher bandwidth. 

Radio waves with lower frequencies are used for farther travel distances, while higher frequencies can carry more data. 5G runs on radio frequencies ranging from below 1 GHz all the way up to very high “millimeter wave” (mmWave), achieving data speeds up to 10Gbps, but is capable of peak data rates up to 20Gbps, in theory. It is currently among the most viable options for facilitating edge computing applications such as autonomous driving, augmented reality or smart cities. 

But is it really fast enough to keep up with the most demanding edge computing applications?

Before we can answer this question, let’s take a look at the most commonly known and most bandwidth-hungry edge computing application: Fully autonomous driving. There are three major types of technologies which are needed for autonomous vehicles to interact with their environment: 

  • Environment information capturing such as cameras, RADAR and LIDAR
  • Direct vehicle-to-everything (V2X) communication based on Wi-Fi protocol 802.11p (DSRC/ITS-G5) or cellular V2X technology (C-V2X)
  • Carrier-based V2X communication based on long-range cellular networks 

It is widely assumed that the latest 5G-network standard will help improve the safety of autonomous cars by providing long-range C-V2X communication, which is a primary requirement.

5G-network

Even at lower level autonomy, connected level 5 cars generate about 25 gigabytes per hour. As complexity of the architecture increases, so does the number of car automation sensors. The combined bandwidth of RADAR, LIDAR, cameras, ultrasonic, etc. can reach up to 40Gbps. As a reference, Tesla’s autopilot system already generates up to 28Gbps with 8 integrated cameras and 12 ultrasonic sensors. The real question, however, is which type of information needs to be transmitted, and if it’s time-critical or not. 

Autonomous driving field tests on German highways such as A9 have shown that even LTE networks can handle latencies of 15 milliseconds. Most time-critical information such as traffic alerts, moving-car data, map updates and critical sensor data have latency requirements of up to 100 milliseconds. Therefore, 5G should be sufficient for enabling more advanced levels of security in autonomous driving, and any other high-speed data transmission at the edge, assuming that the relevant areas are covered with enough network capacity to handle the demand.

Click the buttons below to share this blog post!

Return to the main Blog page




1 Response

Jim Ison
Jim Ison

May 15, 2022

I think 5G will be great for autonomous vehicle to autonomous vehicle real time communications so multiple vehicles can pre-warn other vehicles of their intentions to make everything safer. I also think fleet-wide updates are also a great 5G application. I agree that there is too much data being generated on board an autonomous vehicle with ever more precise sensors for all of that data to be transmitted over 5G. That is why a powerful AI inference system needs to be on board to handle all of the sensor data and make real time decisions because sending all of the data over 5G to a datacenter to get a decision will still take too long for driving. The on-board GPU accelerated AI inference system and 5G will work hand in hand to make large scale Level 5 autonomous driving (and other distributed AI applications) a reality.

Leave a comment

Comments will be approved before showing up.


Also in One Stop Systems Blog

Artificial Intelligence in Navy Ships
Artificial Intelligence in Navy Ships

February 07, 2023

In the last few decades, the United States Navy has opened the door to upgrades and advancements in Artificial Intelligence (AI) technology for its ship systems, command and control infrastructure and decision-making capabilities. The Naval Research Laboratory established the Navy Center for Applied Research in Artificial Intelligence in 1981, and has been in the forefront of emerging technologies for Artificial Intelligence Machine Learning (AI/ML) applications for the United States Navy. These AI technologies are particularly found in two main areas: autonomous/unmanned systems and decision sciences.

Continue Reading

One Stop Systems' Annual Report
The Plan

January 31, 2023

The beginning of the year at a public company is inundated with a myriad of tasks which require a significant amount of time. The first task is writing annual performance reviews for each employee. This includes a review and assessment of achievement of goals from the previous year, and the establishing of goals for the new year. These are required to be written in SMART goal format; Specific, Measurable, Achievable, Relevant and Time based. 

Continue Reading

Which Cooling Method is Best for AI Transportables?
Which Cooling Method is Best for AI Transportables?

January 24, 2023

The most powerful artificial intelligence computing hardware is designed to thrive in a datacenter environment where there is uncapped clean power, near limitless cooling capacity, and a vibration-free environment. The growth of AI use cases in vehicles including automated crop management, autonomous long-haul freight, and military ISR aircraft necessitates the use of datacenter-oriented hardware in vehicles – particularly for initial developments while more customized size, weight, and power (SWaP) optimized embedded platforms are developed. 

Continue Reading

You are now leaving the OSS website