"1 introduction
The formulation of 5g new air interface standard is divided into two stages: rel-15 and rel-16. Rel-15 is mainly designed for the requirements of enhanced mobile broadband (embB) and low delay and high reliability (urllc). The non independent networking version has been completed in December 2017 and the independent networking version has been completed in June 2018 [1-5]. Rel-16 is further enhanced on the basis of the Standard Version formed by rel-15, covering the improvement of basic capabilities, the improvement of mobile broadband capabilities, the expansion of Internet of things services and other aspects.
As the core and key technology of 5g, large-scale antenna technology plays a vital role in meeting the technical requirements of embB, urllc and mmtc (mass machine communication) services. For example, for the embB scenario, the main technical indicators are spectral efficiency, peak rate, energy efficiency, user experience rate, etc. high-order MU-MIMO transmission can obtain very high spectral efficiency. At the same time, with the increase of antenna scale, the effects of inter user interference and noise tend to disappear, and the transmission power required to achieve the same coverage and throughput will also be reduced, Improve energy efficiency. In addition, the large bandwidth in the high frequency band is the key to reach the peak rate. The shaping gain provided by large-scale antenna technology can compensate the path loss in the high frequency band, making the deployment of mobile communication applications in the high frequency band possible. For the urllc scenario, the main technical indicators are delay and reliability. The semi open-loop MIMO transmission scheme enhances the transmission reliability by means of diversity gain. Distributed large-scale antenna or multi Trp transmission technology can further improve the reliability of transmission by dispersing data to multiple transmission points separated geographically. For mmtc scenarios, the main technical indicators are the number of connections and coverage. The beamforming gain of large-scale antenna technology helps to meet the coverage indicators of mmtc scenarios. At the same time, high-order MU-MIMO is also conducive to a significant increase in the number of connections.
Based on the previous accumulation, such as the 3D channel model and scene research started in rel-12 stage, the fd-mimo technology standardization completed in rel-14 stage [6-7], 3GPP completed the standardization of large-scale antenna technology in the first version of NR in rel-15 stage. In this version, the transmission scheme of large-scale antenna technology, channel state information feedback mechanism, reference signal design and beam management are mainly included. For the second version of NR in rel-16 stage, large-scale antenna technology plays an important role in enhancing the ability of mobile broadband. The technical hotspots discussed in the industry include high-resolution channel state information (CSI) feedback, multi Trp transmission, enhanced high-frequency support, enhanced uplink and downlink channel reciprocity support, etc. This paper will first summarize the standards and industrial progress of large-scale antenna technology, then focus on the hot spots of large-scale antenna technology, and look forward to the future development of large-scale antenna technology.
two Summary of technical standards and industrial progress of large-scale antenna
two point one Progress of rel-15 standard
3GPP has carried out many discussions in the Rel-15 stage for large-scale antenna technology, including: high frequency channel modeling, synchronous channel design, control channel design, beam management technology, channel measurement and feedback, multi-point simultaneous interpreting (CoMP), etc.
(1) High frequency channel modeling
Channel model is the basis of system design and performance evaluation. Following the development of three-dimensional spatial channel model (3D SCM) in 2015, 3GPP launched a research project on NR channel modeling in 2016, covering the high frequency band of 6 GHz ~ 100 GHz, and finally formed a technical standard
TS 38.900[8]。 The established high-frequency channel model covers typical deployment scenarios such as indoor offices, shopping malls and downtown, including multiple parameters such as path loss, antenna decoupling, Los probability and spatial consistency, which can effectively support the evaluation of large-scale antenna technology.
(2) Synchronous channel design
The synchronization channel is the key channel for the terminal (UE) to perform cell initial selection, cell synchronization, cell search, uplink access and cell handover. For the high frequency band of NR, the synchronization channel needs to use large-scale antenna technology for beamforming, and use the gain of narrowband beam to achieve sufficient coverage.
The design of synchronization channel is different according to different user states. For connected state (RRC)_ Connected), different base stations can share the beamforming information parameters of their downlink synchronization channel through backhaul, including the number of beams, the position of the synchronization signal in time domain / frequency domain, etc. the source cell can notify the users of the target cell of the synchronization signal parameter information for time-frequency synchronization acquisition during cell handover. For idle state (RRC)_ Idle), UE does not have any prior information of the network before downlink synchronization, so it needs to search all possible synchronization channels. In order to improve the coverage of the synchronization channel, the base station beamforms the synchronization channel to form multiple narrowband beams, and uses beam scanning to achieve full coverage in the cell. The main challenge is how to support different scanning schemes, beam scanning processes in time domain and frequency domain, and downlink synchronization channel design under beam scanning.
(3) Control channel design
Compared with LTE system, 5g NR system supports large bandwidth, and the bandwidth between users in the same cell may be different, and the bandwidth of UE may also be different from that of cell. According to the types of control channels and application scenarios, different MIMO transmission schemes have different advantages and limitations.
The common control channel needs to be sent to all users in the cell. There are usually two MIMO transmission schemes. One is the wide beam scheme, which sends the control channel to the whole cell in a single transmission time-frequency unit to cover all users; The second is the narrow beam scheme, which uses narrow beams pointing to different directions to scan and transmit on different time-frequency resources. At this time, UE needs to monitor the control channels corresponding to different beams in different time-frequency resources, and the detection complexity and delay may increase.
The UE dedicated control channel is sent to the specific ue in the connected state, and its MIMO transmission scheme is related to the user's moving speed. For most users with low moving speed, their position in the cell and angle relative to the base station are relatively fixed. Beamforming of the control channel through a narrow beam can obtain the shaping gain and improve the coverage distance of control signaling. The beamforming steps of the control channel and the data channel may have similar steps. For some ues with high moving speed, their position in the cell and angle relative to the base station change rapidly, so it is difficult to accurately control the beam of the control channel through a narrow beam. A wide beam can be generated through transmission diversity or open-loop transmission to improve the coverage robustness of the control signal.
(4) Channel measurement and feedback
When the number of antennas continues to increase, it is a difficult problem for 5g NR to realize high-precision channel state information (CSI) feedback and maintain low measurement complexity, feedback overhead, power consumption and UE complexity.
According to different duplex modes, the channel measurement and feedback modes are different. For FDD system, because channel reciprocity does not exist or only exists long-term channel reciprocity, channel feedback based on downlink measurement is still the main CSI acquisition scheme. Key factors to be considered include: number of csi-rs ports, new air interface csi-rs design, high port feedback codebook design and enhancement, high-precision CSI feedback scheme, explicit feedback, etc. 5g NR rel-15 standardizes two types of codebooks, type I and type II codebooks. Type I is a conventional precision codebook, which is used to support single-user MIMO and multi-user MIMO transmission. Type II is a high-precision codebook, which is mainly used to support multi-user MIMO transmission and improve the spectral efficiency of the system. For TDD system, due to the reciprocity of uplink and downlink channels, the downlink channel information can be obtained through uplink channel measurement. The key factors considered are: how to use a few uplink transmitting antennas to obtain the channel of most downlink receiving antennas, interference measurement enhancement, CQI calculation enhancement, etc.
(5) multi point simultaneous interpreting (CoMP)
The LTE system starts with Rel-11, and supports multi-point cooperative transmission, including simultaneous interpreting (JP), cooperative scheduling / CS/CB and dynamic point silence (DPB/DPM). At the same time point, the data of UE can be sent by one transmission point (TRP) or jointly by multiple TRPs. The channel state information of different TRPs is fed back by multiple CSI processes.
Due to the expansion of antenna scale, the beam of NR system becomes narrower, which can more accurately adjust the transmission angle and more flexible interference cancellation. Therefore, the application of comp will be more flexible and complex. The main technical scheme is enhanced comp transmission scheme. For example, multi-point incoherent joint transmission (nc-jt), UE receives multi stream data transmission at the same time point, and different data streams are sent from different TRPs; Under the condition of high-density networking, CSI feedback supports more transmission points and more interference assumptions.
two point two Industrial progress
With the research and standard promotion of large-scale antenna technology, the industry has rapidly developed the base station prototype and products supporting large-scale antenna technology, and conducted technical tests. Taking the 5g technology test in China as an example, in the first stage (2016), manufacturers developed large-scale antenna arrays of 128 yuan or more, among which Datang Telecom Group developed the industry's largest 256 yuan large-scale antenna system, which can support the cell peak rate of 4 GB / S based on 100 MHz bandwidth. In the second stage (2017), for the pre commercial verification requirements of operators, manufacturers have developed 64 channel 192 yuan large-scale antenna arrays. Among them, the base station products developed by Datang Telecom Group support 200 MHz bandwidth. In the test, they can support 28 streams. The single user peak rate reached 1.6 GB / s, and the cell peak rate exceeded 10 GB / s. Since the third phase (2018), the pre commercial and commercial products supporting large-scale antenna technology based on non independent Networking (NSA) and independent Networking (SA) systems have started system level verification. The preliminary verification shows that for the 100 MHz bandwidth 256qam modulation mode, the 8 downlink terminals can support 16 streams to reach the peak rate of 6 GB / s cells, 16 downlink terminals can support 32 streams to reach the peak rate of 10 GB / s cell.
three Enhanced large-scale antenna technology
The multi antenna information theory proves that [9], when multiple antennas are used at both ends of the wireless communication link, the channel capacity of the communication system will far exceed the information transmission capacity limit of the traditional single antenna system. This theory provides a solid theoretical basis for large-scale antenna technology and shows its broad application prospect in high-speed wireless access system. The application scenarios of large-scale antenna technology discussed in the industry include centralized coverage, distributed coverage, high-rise buildings, heterogeneous network scenarios, indoor and outdoor hotspots, suburbs, wireless return links, etc. [10]. The frequency band directly determines the size of the antenna system. In scenes requiring wide area coverage, large-scale antenna technology tends to be in the low frequency band below 6 GHz; In scenarios such as hot spot coverage or return link, large-scale antenna technology tends to be in the high frequency band above 6 GHz. The application of high frequency band is very beneficial to the miniaturization of large-scale antenna array and actual network deployment, and in the high frequency band, the high beam gain provided by large-scale antenna system is also needed to compensate for the influence of non ideal factors in the propagation environment. Therefore, 5g NR supports the basic functions of large-scale antenna, and the large-scale antenna technology will be further enhanced in the enhanced 5g system. The following focuses on some technical hotspots of large-scale antenna technology for 5g enhancement, including high-resolution channel state information feedback, multi Trp / panel transmission, beam management, enhanced dl-ul channel reciprocity support, etc.
three point one High resolution channel state information feedback
In terms of network performance, CSI feedback of type II is the most potential means to improve the performance of large-scale antennas. In 5g NR, linear combination of L = 2, 3 and 4 beams is supported, and only codebooks with rank 1 and 2 are supported. Compared with the ideal CSI feedback performance, the type II codebook supported by 5g NR still has a performance gap. At the same time, due to the huge feedback overhead of type II codebook, it also increases the complexity of system design when reporting CSI.
In the enhanced 5g system, on the one hand, the codebook performance needs to be improved, on the other hand, the feedback overhead needs to be reduced. In order to improve the codebook performance, the number of beams for linear merging can be further increased. For example, l > 4 is supported, and the feedback overhead can be dispersed in multiple reports by means of differential reporting. In addition, codebooks with rank 3 and 4 need to be further supported. This codebook can be applied to ue with more than 8 receiving antennas and scenarios with light network load to improve throughput. In order to reduce the feedback overhead, the frequency domain compression scheme can be used to compress the codebook parameters of the subband. Frequency domain compression scheme refers to using the correlation between different subbands of CSI calculated by UE to sparse express the CSI calculated by UE through a set of orthogonal bases in frequency domain, and UE quantizes and feeds back the coefficients after sparse expression. Depending on the scenario and channel conditions, the system can save up to 60% of the feedback overhead after frequency compression. Fig. 1 and Fig. 2 respectively show the performance evaluation results of different coefficient quantization schemes. Alt1 refers to that the amplitude of the coefficient is quantized by 3 bits, alt2a refers to that the amplitude of the coefficient is quantized by difference, alt3 refers to that the amplitude of the coefficient is decomposed into the Kronecker product of spatial coefficient and frequency domain coefficient and quantized, and the phase of the coefficient is quantized by 2 bits and 3 bits respectively, α Represents the proportion of the coefficient phase quantized with 3 bits. The base station in the simulation environment is 128 antennas and the terminal is 2 antennas. See Table 1 for other specific simulation conditions. In Figures 1 and 2, the ordinate "relative performance" shows the ratio of the performance of the compression scheme to that of the uncompressed scheme, and the abscissa "overhead (bit)" shows the feedback overhead of the compression scheme, while the feedback overhead of the uncompressed scheme is 455 bits. From the evaluation results, it can be seen that the compression scheme greatly reduces the feedback overhead, but the impact on the performance can be ignored.
three point two Multi Trp transmission
For multi Trp transmission, the standardization supporting nc-jt scheme is not displayed in 5g NR, so it needs to continue research and standardization in 5g enhanced system. In order to enhance multi Trp transmission, it is necessary to further increase the number of cooperative TRPs, such as extending to more than 2. At the same time, the transmission scheme can be designed in combination with a variety of application scenarios, such as coherent / incoherent transmission, uplink and downlink cooperation, indoor and outdoor scenarios, joint transmission (JT) / cooperative beamforming (CB) / transmission point selection (DPS), etc.
In the 5g enhanced system, it is necessary to design around the transmission scheme of multi Trp and CSI reporting mode. In the design of CSI reporting mode, both beam reporting and CSI feedback need to be enhanced. In terms of transmission scheme, two-level control signaling can be used to support incoherent JT. In addition, multi Trp transmission of control channel can also be studied
Our other product:
