Crucial to successful network deployment, drive testing of 5G NR networks is a multifaceted pursuit.
In our ongoing series of blog posts, we’ve already looked at how technical disruptions are changing the way 5G NR testing must be done, we’ve discussed why it remains relevant for 5G NR and how it must be transformed, and we’ve explored ways of rethinking TEMS probes.
In this, the fourth part of the series – which ties in with our detailed white paper, Initial 5G NR Drive Testing with Infovista – we turn our attention to multiple-input, multiple-output (MIMO), specifically massive MIMO (mMIMO) and 3D beamforming systems.
There’s lots to consider when testing mMIMO /3D beamforming, and we’ll cover aspects relating to:
- Performance evaluation;
- Troubleshooting; and
But first, a quick look at the background.
Background to mMIMO/3D beamforming
What are its advantages?
Well, by using more antennas intelligently, network capacity, coverage and throughput per user can be improved. That is, more spatial data streams can significantly increase spectral efficiency (such as with multi-user mMIMO), allowing more bits to be transmitted per Hertz.
At the same time, smart beamforming techniques can extend the reach of base stations by focusing RF energy in specific directions on the downlink, similarly enabling the base station receiver to capture energy from a specific direction with less noise and interference on the uplink. This enables higher throughput per user, as well as increased coverage, especially at the cell fringe.
In addition, harnessing mMIMO technologies using 2D antenna arrays at the base station for 3D beamforming enables the use of higher frequency bands of mid-band (3GHz – 6GHz) and above 6GHz spectrums.
However, accurate and timely channel knowledge is essential to realizing the full benefits of 3D beamforming. In order to achieve this, optimized design for fast reciprocity-based time division duplex (TDD) mMIMO makes use of the self-contained slot structure and enhanced reference signals to support much faster and more accurate channel feedback. Note, though, that these optimization algorithms are device vendor-specific. Consequently, different performance gains are expected on different devices for the same network scenario.
This brings us to the main testing aspects that you should consider to evaluate, troubleshoot and optimize the performance gains of mMIMO/3D beamforming.
1. Respond to complex antenna patterns with thorough planning
Planning tools for mMIMO/3D beamforming systems need to cope with complex antenna patterns, which must be modeled, validated and tuned using drive testing data collected within best suited environments (geography, demography, weather). They’re expected to trigger MU-mMIMO and/or 3D beamforming, depending on the targeted spectrum (low/mid/high frequency. Therefore, more precise, automatically and remotely controlled drive testing campaigns are needed.
Taking care of such a thorough mMIMO network planning before testing it helps avoid impactful consequences of missed or incorrect engineering design aspects. Risks include overspending on network CAPEX, failure to capture revenues where the traffic is and your radio network performing poorly due to lack of coverage and/or capacity, leading to network instability and low QoS.
All of these could result in a much longer testing phase.
2. Test horizontally and vertically
3D beamforming requires 3D testing in order to evaluate the performance (such as coverage and interference) of the beams that are planned and deployed in both the horizontal and vertical plane.
Every beam needs to be measured, evaluated and visualized in 3D, which requires drone-based testing.
3. Understand the phases of beam management
The beam-centric characteristics of 5G NR require beam-centric coverage and interference evaluation using device-based measurements for beam-finding, tracking, acquisition and switching during the process of aligning transmitter and receiver beams. This process is governed by beam management, which consists of two phases: initial access and beam switching.
The first phase refers to establishing the radio link to idle users, as in devices that aren’t actively transmitting. The switching phase refers to users who are already connected to the network. In this phase, procedures involve handover, path selection and radio link failure recovery.
The 5G NR specifications include a basic set of beam-related procedures for the control of multiple beams that need to be tested and evaluated, especially because beam management is vendor-specific, meaning different performances are likely. Performance evaluation needs to address beam sweeping (the beam set covering an area), measurement (quality of the received signal at gNB/UE), determination (best beam selection) and reporting (UE beam quality reports to the RAN).
4. Take account of device-level variations
Since beam management is to a large extent vendor specific, detailed analysis of the beam management performance evaluation at device-level is required. This is especially important during beam mobility scenarios, both with intra- and inter-site beam switching/HO . In addition, in the mmWave case, beam mobility at device-level needs to be evaluated both in idle and connected mode with various blockage conditions created by different building types and structures, and/or different vehicles types.
5. Always be benchmarking
Understanding the performance benefits of NR, both non-standalone (NSA) and standalone (SA) mode, versus legacy LTE, requires continuous benchmarking of coverage and throughput within the context of spectral efficiency. NR performance is expected to be achieved with better spectral efficiency than LTE, especially for NSA mode, although coverage and throughput performance might be similar between the two networks in some specific scenarios. This enables higher capacity.
6. Evaluate DL and UL balance below 6GHz
Evaluate downlink and uplink (DL/UL) balance on the sub-6GHz spectrum. This is needed because such a wavelength limits antenna array size in user equipment (UE), so it will transmit far fewer and much wider beams than downlink.
7. Evaluate beam interference effects on coverage and performance
The beam-centric characteristics enable the coordination of beams from several transmission points. This increases the complexity at UE-level, specifically the UE’s sensitivity to interference occurrence. So, you must evaluate to what extent (if any) coverage and performance (i.e. throughput) is affected by beam interference.
Testing of these aspects requires the measurement of new Information Elements (IE) per individual beam, as well as decoding a series of generally known types of reference signals (RS) and IE/KPIs. You should also monitor the re-use of signals from the same sector in different beams. Testing and evaluation consist of two main phases that require several main IE/KPIs, irrespective of network access and reference signals.
Increased capacity, higher throughput per user, and 3D and increased cell fringe coverage in 5G NR are mainly ensured by the evolution of MU – mMIMO/ 3D beamforming systems. However, for an optimized benefit and functionality, both careful planning and prediction tuning as well as testing is needed.
In this blog we discussed the 7 main considerations. But you can find more details and measurement results in our white paper, Initial 5G NR Drive Testing with Infovista.