When it comes to making the most out of a radio channel, 5G NR goes to even greater lengths than LTE. There are lots and lots of mechanisms for the base station and the UE to analyze current channel conditions and, based on the results, optimize their transmitter and receiver settings accordingly. Quite a number of posts over at Sharetechnote go into the details of this and it’s easy to loose sight of the basic mechanisms at work. So I thought I’d do a quick post on how Channel Measurements for MIMO and beamforming in 5G NR works in principle.
To make the most out of a radio channel, i.e. transfer multiple data streams simultaneously to one or more devices means the transmitters and receivers need to be aware of their current channel conditions. In other words:
- Devices need to analyze how many distinguishable radio paths on the channel exist for them at their current location due to reflection of radio waves from various objects around them.
- In addition, they need to be able to distinguish between the actual signal for them and other radio energy not intended for them, i.e. signal noise. This noise can be caused by transmissions to other devices in the same cell and and transmissions from neighboring cells, i.e. neighbor cell interference.
These things can be done by the receivers by analyzing transmissions of which they know how they look like, i.e. what was being sent. Like LTE, 5G NR uses a resources grid over the air, i.e. many sub-carriers on the frequency axis with a bandwidth of e.g. 30 kHz, on which symbols are transmitted that encode one or more bits per transmission step. Symbols that carry known data for channel analysis are referred to as Reference Signals (RS).
In the downlink direction, 5G NR uses Primary and Secondary Synchronization Signals (PSS and SSS) as part of regular broadcasts that in addition contain system information so the UE can find the network in the first place and keep a lock on it. In case beamforming is used in the cell, these regular broadcasts are beamformed in different directions and frequently repeated. These transmissions are referred to as Synchronization Signal Blocks (SSBs) and a typical setup in the 3.5 GHz range uses 6-8 beams for the synchronization and broadcast information. All broadcast beams with system information are repeated frequently, 20 milliseconds is a typical value in practice. In other words, the UE can analyze very quickly which beam is best for reception. It can then tell the network during connection establishment which beams are received with which signal strength and the network can then use this information for initial beamforming.
Once the UE has an active bi-directional channel to the network, the beam to the UE can be further refined and steered based on a number of mechanisms of which I would like to quickly introduce two:
Mechanism 1: To give the UE further opportunities to measure the channel, the network can insert Channel State Indication Reference Signals (CSI-RS) into the downlink data transmission to that UE. That means that not all symbols of a Resource Block assigned to a UE are used for user data but some are used for reference symbols whose content is known by the UE. After analyzing how the reference signals were received at the different antennas at each UE (e.g. signal strength, delay, distortion of the signal due to noise, etc.) the UE can then send feedback to the network. In practice there are several ways to do this: One of them is the Precoding Matrix Indicator (PMI) which the network can instruct the UE to periodically return. The PMI points to a set of values the network can then use to control the delay and signal strength of downlink transmissions over different antennas.
Mechanism 2: Another way that can be used alternatively or in addition to mechanism 1 is for the base station to analyze transmissions of UEs in the uplink direction. To do this, the UE must send reference signals in the uplink direction whose content is known to the base station. These signals are referred to as Sounding Reference Signals (SRS) and are typically configured during RRC connection establishment. Especially in TDD systems where uplink and downlink are on the same channel, the Sounding Reference Signals in the uplink direction can help the base station to control transmissions in the downlink direction.
So there are quite a number of different ways to estimate channel conditions. In essence it comes down to transmitting known symbols in the uplink and downlink direction, to analyzing them in the UE and in the base station and to adapt future transmissions accordingly.