Massive beamforming in 5G radio access
How can you get all users in high rise buildings >200 Mbps throughput wherever they are, and reduce network energy consumption at the same time? Through advanced antennas, flexible, user-specific beamforming and a new mobility algorithm.
The key is to use spectrum at high carrier frequencies, and to deploy advanced antennas to combat the challenging propagation conditions at these frequencies. We’d like to share with you the facts of our solution, which we demonstrated at the Mobile World Congress recently.
Multi-antenna technologies have always been an integral part of LTE. As 5G research is picking up speed, it’s time to take advanced antenna systems to the next level. Clearly, advanced antennas will be a crucial part of 5G radio access, and one important reason for the soaring interest is that 5G radio accesses initially will be deployed at high carrier frequencies.
High carrier frequencies present challenges: At high frequencies, the propagation is more hostile. The free-space propagation loss is higher, since the receive antennas get smaller with higher frequency. The diffraction losses are higher because the radio signals do not “bend” around corners to the same extent at high carrier frequencies. Finally, the wall penetration losses are higher.
Higher frequencies, however, also offer significant opportunities: As the carrier frequency gets higher, the antenna elements get smaller. With this, it becomes possible to pack more elements into a smaller antenna. For example, a state-of-the-art antenna for 2.6GHz is roughly one meter tall, and contains 20 elements. At 15GHz, it is possible to design an antenna with 200 elements that is only 5 cm wide and 20 cm tall.
With more antenna elements, it becomes possible to steer the transmission towards the intended receiver. Since we are concentrating the transmission in a certain direction, coverage is significantly improved.
With more antenna elements, the beams get narrower. It then becomes vital to transmit the signal in the appropriate direction, to maximize the received signal energy at the mobile. Such a scheme is called user-specific beamforming. In our solution, we steer the transmissions in both the vertical and horizontal dimension, specifically to each individual user. Since one cell may serve hundreds of users, the beam direction may change several times per millisecond.
When we have installed an antenna that can provide massive beamforming, we quite naturally want to use this antenna for as many transmissions as possible. The term ‘massive beamforming’ alludes to the high number of antenna elements per antenna, in our case >100. In academia, there is a concept called massive MIMO, which usually implies that we transmit to several users at the same time, i.e., multi-user MIMO. Multi-user MIMO will be included in future versions of our concept.
However, some transmissions do need to reach several users. Examples are synchronization signals and system information needed for initial system access. In today’s cellular systems, such signals are common. Since we cannot use beamforming to e.g. overcome challenging propagation conditions for broadcast signals, we need to minimize that type of signal. We call the approach ultra-lean design. It includes, for example, removing cell-specific reference signals and replacing broadcast with dedicated signaling.
Ultra-lean design has other important advantages, in addition to maximizing use of beamforming, e.g. for significantly improved coverage:
It reduces inter-cell interference from always-on signals.
It enables long DTX periods in the base stations, leading to large savings in base station energy consumption
The potential gains related to energy consumption are remarkable. Existing standards specify that base stations transmit signals – at least – every millisecond. For 5G, we propose to extend the sleep mode period, with no required transmission, to some 100 milliseconds. This does not only increases the potential for sleep mode, but also for longer sleep-mode periods enabling more efficient energy saving.
Support for massive beamforming, operation at high carrier frequencies and ultra-lean design were included in the first version of the advanced antenna concept we developed in 2014. Quite early in the process, we also wanted to develop an understanding of what kind of performance you could actually expect in a system where massive beamforming is applied. Since there are no commonly used propagation models for 15GHz carrier frequencies, we rely on a ray-tracing type of propagation model for a city. It’s designed to resemble the central parts of a major Asian city, like central Tokyo or central Seoul. The inner part of the evaluated area contains many tall buildings - up to 150 m - surrounded by an area with lower buildings. The city is covered by a dense outdoor macro network. The propagation model has been benchmarked towards measurements of similar deployments in similar city environments.
We evaluated performance for this network deployment. In line with the Ericsson strategic forecast for 2020, we assume an average subscriber in the city uses 32GB per month. With reasonable assumptions of population density, the system needs to handle 1600Mbps/km^2. Today, a typical solution in such a city is to use 40MHz LTE FDD, deployed in a tight site grid. While providing very good performance today, such a solution will be inadequate in 2020, as shown in the picture below. “Red” means that the user throughput is below 10Mbps.
To improve performance, we deploy a 5G system at 15GHz with 100MHz TDD, using massive beamforming with an antenna with 200 elements. Remember, such an antenna is still quite small: 5 cm wide and 20 cm tall. The system is designed to enable very flexible UE beamforming, and implements an ultra-lean design of its control channels. We call the system NX.
With NX deployed on the same sites, we apply massive beamforming using antennas with 200 elements (today 20 elements) to get this performance:
The introduction of massive beamforming leads to significantly better performance. The maximum antenna gain is 28dBi, meaning that it radiates 28dB (630 times) higher energy in the desired direction than an omni-directional antenna would. The large maximum antenna gain in combination with user-specific beamforming makes it possible to overcome the challenging propagation conditions at 15GHz.
To boost performance, we apply carrier aggregation between the 2.6 and 15GHz layers. The users with really challenging propagation conditions can then be handled by the 2.6 layer. Of course, the addition of more spectrum also helps:
The darker green color means that user throughput in these buildings exceeds 200Mbps.
To conclude, we see that massive beamforming provides great gains, also in very challenging propagation conditions. With a tight outdoor deployment, it seems indeed possible to provide very good indoor performance. Carrier aggregation between higher and lower frequencies also shows beneficial.
Our work now continues to detail the concept and to take it further by including, among other things, multi-user MIMO.
Claes Tidestav, Senior Specialist in Radio Network Algorithms