How Bell Labs is Revolutionizing the Mobile Technology Game That They Invented

They are making 5G happen.

Has Bell Labs created anything noteworthy recently? originally appeared on Quora, the place to gain and share knowledge, empowering people to learn from others and better understand the world. You can follow Quora on Twitter, Facebook, and Google Plus.

Exactly 70 years ago (in 1947), Bell Labs brought us transistor. Scratch that, actually. Exactly 70 years ago, Bell Labs brought us cellular telephony. It took a while for the technology to catch up, and Bell Labs wasn’t the one to win the race (Motorola did, and they made sure Bell Labs was the first to hear all about it), but that’s how it all started. Now, it’s Bell Labs again: they are making 5G happen.

The story begins in a rather academic setting, with a scientific paper proposing a theoretical concept. Ring a bell?

In 2010, Tom Marzetta’s article titled Noncooperative cellular wireless with unlimited numbers of base station antennas appeared in IEEE Transactions on Wireless Communications. Now, treatises of multiple antenna systems were nothing new in theory and practice, neither in Bell Labs (more than a decade before this paper, Gerry Foschini published his highly influential work On limits of wireless communications in a fading environment when using multiple antennas, followed by his presentation of V-BLAST, Vertical-Bell Laboratories Layered Space-Time), nor elsewhere (the concept of MIMO - multiple input multiple output - in wireless communications was the fundamental idea of Stanford’s Arogyaswami Paulraj five years before Foschini’s work). So, what was so special about this 2010 paper? Obviously it wasn’t very practical, as there are a few practical problems in manufacturing an unlimited number of antennas, namely the finite time and resources we have.

Before answering this question, let’s define a few terms and see what this MIMO story is all about. So, opposed to SISO (single input single output) systems, MIMO is all about multiple antennas on a base station, both transmitting and receiving from the users in the cell the base station is covering. Now, if you have a single antenna transmitting to multiple users, you need to have a way to have everyone get their own content, and not someone else’s. This is called multiplexing, and two elementary ways to conduct it are in time and frequency domain: either use different frequencies for different users, or transmit to different users at different time slots (or both). Now, if you add more antennas at the base station, you get one more option, spatial multiplexing: you can have multiple base station antennas transmitting parts of the signal to the user in question and have those signals add up to the real thing only at the location of this user (elsewhere it does not). Furthermore, with more antennas, the signals can be transmitted over different paths to the user, and hence cancel the effects of fading.

Now, all of this was known for MIMO systems in general ever since Paulraj introduced them and Foschini elaborated on their applicability. What about this outrageous infinite antennas idea? It did show that the use of a large number of antennas in MIMO base station systems brings significant improvement for the throughput. Sure, it sounds like a greedy algorithm application: just throw in more antennas. It does make sense, but on the other hand, the existing concepts of point-to-point MIMO and multi-user MIMO aren’t very scalable due to significant time (training, coding/decoding, channel state information collection time) and space (line-of-sight) related challenges.

Massive MIMO, as proposed by Marzetta, has superb scaling properties: the channel state information acquisition time is independent of the number of base station antennas and simple linear precoding has no significant time effects and at the same time, it approaches the optimum as the number of antennas grows.

Simple precoding is important, because at small numbers of antennas, complicated precoding schemes are used to get the channel capacity as near to the optimum as possible (for instance, dirty paper coding is sexy and fantastic, but computationally demanding for a larger number of antennas). The channel state information acquisition is also crucial, if you recall the story of spatial multiplexing mentioned earlier: send the signal from various antennas so it adds to the desired signal at the user’s location. In order to know what will happen to the components sent from different antennas to the user on their way to their destination, we need to know the characteristics of the channel between each one of the base station antennas and user’s antenna(s): the attenuation and phase shift.

Massive MIMO concept [1]

It took a few years for Massive MIMO to get its practical implementation: last year the testbed in Bristol was presented as a great testing platform for Massive MIMO, and more importantly, the major telecom companies around the world started manufacturing their Massive MIMO base stations: Huawei and ZTE, followed by Ericsson and Nokia recognized the importance of getting a Massive MIMO product to the market. Even though the current standard does not allow Massive MIMO to unleash its full potential, we ought to expect Massive MIMO as the fundamental underlying concept in 5G, together with mmWave and small cells. Huawei Helps China Mobile Shanghai Deploy World's First Massive MIMO Solution.

So… yes, Bell Labs enabled 5G. Will it help its owners, Nokia, in making Massive MIMO a reality, or will the new Motorola-likes repeat history? Only time will tell.



This question originally appeared on Quora. More questions on Quora:

* Bell Labs: What was it like to work at Bell Labs?

* Mobile Cellular Technology: Can WiFi fully replace 4G/5G?

* Technology Trends: What is Apple HomePod and why is it important?

Photo Credit: exdez/Getty Images