Blog

2 August 2022, Tomas Zvolensky, PM

Multi-User MIMO

Multi-User Multiple Input Multiple output (MU-MIMO) systems are all the rage these days in communication systems of all levels, be it consumer or carrier class environment. In this blog you will learn about:

Difference between SU-MIMO and MU-MIMO system
MU-MIMO system requirements
Role of antenna arrays in MU-MIMO
MU-MIMO upload & download

SU-MIMO vs. MU-MIMO

In our introduction to MIMO we elaborated on Single User MIMO (SU-MIMO), where the pre- and post-processing happens on both sides of a wireless link - the Access Point (AP) and Customer Premises Equipment (CPE). SU-MIMO systems work well in scattering-rich environments, but poorly in a Line of Sight (LOS) scenario with little to no multipath propagation because the signals arriving at different AP antennas are very similar. Because of this, the AP cannot separate the CPE signals from one another.

In a MU-MIMO system, the pre- and post-processing only happens on the side of the AP which communicates with multiple independent CPEs at the same time.

MU-MIMO system requirements

There are two major differences between SU-MIMO and MU-MIMO on system level:

Channel State Information (CSI) or sounding - before the link with each CPE can be established, AP needs to know what the propagation channel to each CPE looks like. So the AP sends test data to every CPE. Based on the predetermined knowledge of the test data, the CPE can calculate the channel response and send it back to the AP, which then uses it during DL for correct operation.

CPEs need to be synchronized for correct UL functioning. During UL, the signals from all CPEs must arrive at the AP at the same time. Since they are typically in different locations, the CPEs need to know what delay to introduce such that their signals arrive at the AP simultaneously during UL.

Once the MU-MIMO system is set up, the links can be established. MU-MIMO systems typically use antenna arrays to maximize the Signal to Noise Radio (SNR) of signal each CPE receives. If you want to know more about antenna arrays (AA), check our short blog post.

Depending on the number of CPEs, there is a corresponding number of independent data streams the AP works with. For each data stream, the AA is configured such that the main beam of the radiation pattern is aimed at the corresponding CPE. So the individual setup of AA is different for each CPE. Once all AA setups for each CPE are known, the final radiation pattern is a sum of the individual ones.

Role of antenna arrays in MU-MIMO

Because of the antenna array size, there is a limit to how many CPEs can be served at the same time - minimum beam width size limits how close the served CPEs can be. If CPEs are too close, they are covered by the same beam and therefore cannot be separated from each other in a logical sense. To be able to do so, the beam width would have to shrink even more, which is possible by increasing the AA size.

There are two major ways to determine which CPEs can be grouped and served at the same time. First is to search for the maximum achievable SNR among all CPEs and picking those that have the best SNR.

Second is to search for the opposite - finding directions to which the nulls are steered to aim them at the CPEs that are not part of the served MU-MIMO group.

MU-MIMO upload & download

It is important to understand that the transmission of the signals to all CPEs happens at the same time, and at the same frequency channel. During the download, each CPE receives all the signals containing all the data streams involved. Thanks to the signal processing techniques and the knowledge of the propagation channel characteristics between AP and each CPE, the CPEs can pick the data stream meant for them.

During the upload, synchronized CPEs send their data to the AP, which again, thanks to the knowledge of propagation channel characteristics associated with each CPE and signal processing techniques then sorts through the signals received and separates them into individual data streams.