This newly published book belongs close at hand for every engineer in advanced wireless. Marzetta invented it at Bell Labs, so I expected a fine book. Marzetta and co-authors Larsson, Yang, and Ngo did an extraordinary job. The book is admirably clear, short, and definitive. They answer the key questions: what it is, why it works, and how to design the systems. The last chapter reviews the problems still to be solved. It's only 160 pages (and 50 more in appendices) but all the main topics are addressed. For depth on a particular topic, the authors point you to the original research. The book is written for engineers; some parts are hard going for a layman.
The Resources list at the end references the seminal works by Paulraj, Foschini, Alamouti, Goldsmith, and the authors, as well as the 150 other works that have defined the field.
Tom Marzetta of Bell Labs has been called the "Father of Massive MIMO." His 2010 paper, Noncooperative Cellular Wireless with Unlimited Numbers of Base Station Antennas, has been cited over 1700 times. He and Bell Labs colleague Gerry Foschini have been working on MIMO since the 1990's and made many contributions. For those who want to understand in depth, I've included the abstract of that paper and others below.
In 2014, Tom told me he thought it would take several more years for practical systems. Masayoshi Son of Softbank was unwilling to wait and launched the first commercial deployment in September, 2016. Softbank is installing 100 systems across 43 cities in Japan. Softbank's early results, from five cities, show a 5X to 10X improvement in the same spectrum. They use 128 antennas. Some of the antennas are used for "beamforming," which is proving crucial for deployments. Many top engineers expect a 50X improvement from MIMO in the coming years. Remarkably, the increased performance does not require significantly more power.
Marzetta writes, "Massive MIMO is the most promising technology available to address the ever increasing demand for wireless throughput:
• Orders of magnitude spectral efficiency gains over LTE - large numbers of users communicate simultaneously over entire allotted spectrum through elementary multiplexing signal processing
• Uniformly excellent service throughout the cell - regardless of location relative to base station
• Drastically reduced radiated power
• Simple and scalable design - employs measured channel characteristics rather than assumed channel characteristics
• Naturally green technology - superior energy efficiency
128 antennas, 6X-10X improvement in throughput to ?1.5 gigabits. The fearless Masayoshi Son wants to change the world again, this time by launching 5G. They have deployed 100 cell sites in 43 cities, with many in Tokyo. He may be a year ahead of anyone else. Masa-san wasn't patient enough to wait. (I'm mostly working in Google translation from the Japanese, included below.)
Softbank subsidiary Wireless City Networks is taking charge. They may be using the ZTE 128 antenna unit demoed last year. Cell siting is particularly difficult in Japan and MIMO is the best way to get much more out of your existing network and spectrum. On the other hand, Softbank controlled Sprint has 120 MHz of unused spectrum. Sprint's logical path forward begins with carrier aggregation putting that spectrum to use.
Tom Keathley of AT&T approves. No one believed Arogyaswami Paulraj in the 1990's when he claimed MIMO (his invention) would one day lead to a 100x improvement in wireless capacity. That day is now close, with Ericsson making the first announcement of commercial Massive MIMO.
"We must have MU-MIMO," Sanyogita Shamsunder of Verizon said this spring. "Massive-MIMO, also known as 3D MIMO, is an important milestone in China Mobile's technology roadmap," Huang Yuhong of China Mobile says in the pr below. Keathley adds a comment from AT&T.
Many of us have seen how well four antennas work in 4x4 Wi-Fi, raising theoretical speeds over a gigabit and real world tests at 500 megabits. Four antenna MIMO is a crucial enabler of gigabit LTE, about to start deploying. Add more antennas and performance continues to improve. See 1000% MU MIMO gain in gigabit tests by Universities, Facebook.
Millimeter wave is exciting but the telcos are putting MIMO multi-antennas on the front burner. Verizon, NTT and other carriers presented MU MIMO as crucial to their plans at the remarkable Brooklyn 5G event. Sanyogita Shamsunder (quoted in title, pictured) leads their 5G program and made clear in her slide that Verizon 5G will be MU MIMO, not just the publicized millimeter wave.
Tom Keathley of AT&T had a similar comment. "The 5G Industry expects Massive MIMO." He noted the importance of "Sub 6 GHz for wide area coverage with improved spectral efficiency." The Bells seem to be reaching a similar conclusion as NTT. Seizo Onoe, NTT CTO, startled last year's Brooklyn 5G Conference by saying that NTT's 5G efforts mostly will be below 6 GHz for years. High frequencies will be important, Onoe told us, but mostly after 2022-2023.
This year, Verizon and other carriers confirmed that MU MIMO, with many antennas, is crucial to their plans.
The Facebook PR will get everyone noticing MU-MIMO. The top wireless engineers yawned. They have known for years that you could get extraordinary results from adding antennas and multiple streams. Bristol used 128 small antennas in about eight feet by four feet for a 12X efficiency gain. Facebook used 96 antennas (pictured) for what they considered a 10X gain. These are very early units. Better software and dedicated processors will take speeds higher; Facebook promises ~40% improvement. (Many details and excellent video below.)
20 MHz of spectrum could deliver a gigabit. Early LTE networks, like AT&T, get ~100 megabits. More recent LTE-A networks in Turkey, the Philippines, and many other countries use 60 MHZ and will get speeds well into the gigabits.
An important caveat: MIMO works because the signals from each antenna can be separated by how they bounce off walls, etc. That means that the MIMO gain is much less in some locations. Line of sight is best for today's systems but not ideal for MIMO. These are ideal figures; at the edge of a cell you get much less.
Arogyaswamo Paulraj invented MIMO in 1993. He predicted twenty years ago that you would one day be able to increase efficiency 100X.
Page 20 of 22