:::

Research on 5G Mobile Networks and Systems

:::

Introduction

Next generation mobile cellular systems are envisaged to have much higher system capabilities than the current IMT-advanced systems, which will support a widely diversified usage scenarios of mobile access in the year 2020 and beyond including very high data rate communication in a very crowded environment, ultra-low latency and reliable communication, massive machine-type communication, and others. It is foreseen that the new systems will have flexible system architecture to adapt to new usage scenarios which may not yet be identified today and have a wide variety sets of capabilities to meet the different system requirements in different usage scenarios.  
    Large-scale antenna (LSA) technology in above 6GHz bands has been widely recognized as one the most promising technologies to enable the next generation systems to meet the requirement of very high data rate communication in a very crowded environment.  LSA technology has been a topic of intense research around the world. Nevertheless, the currently popularized spatial-domain processed LSA technology, where spatial channels of transmit-receive antenna pairs are treated separately, incurs huge system complexity and system overhead in terms of number of spatial-channels to be processed, complexity of applying MU-MIMO, number of reference signals required, number of spatial-channel estimations and feedback. In addition, the channel frequency selectivity inherent in the spatial-domain processed LSA technology may limit the bandwidth usable by a UE, and that limits the users’ achievable data rate.  Lastly, due to the huge downlink and uplink overhead for acquiring channel state information, the technology is only applicable to TDD systems where the channel reciprocity could be exploited. As a consequence, a new type of LSA technology has to be devised before the LSA technology can be applied to real systems to meet 5G requirements in a cost effective way.
   In this research project, new radio access technologies (RATs) with a beam-domain processed LSA array are investigated. In the beam-domain processed LSA technology, the spatial SU/MU-MIMO channels are transformed into beam-channels, each with a narrow beam-width determined by the size and geometry of the array.  In most propagation scenarios, the number of beam-channels seen by a UE is much smaller than that of spatial channels so that the number of reference signals and channel processing is reduced largely, and that significantly reduces the system complexity and overhead. Furthermore, by compensating the delays between beams of a UE before signal processing, the channel frequency selectivity can be resolved in the beam domain and hence improving the user data rate by allowing it to use all the available bandwidth.
The design goal of the new RATs is to support user experienced data rate of 100Mbit/s – 1Gbit/s, peak data rate of 10 – 50 Gbit/s, mobility of up to 500km/h, latency of 1 ms, and 5 times of spectral efficiency and 50-100 times of energy efficiency than the current IMT-advanced systems. To achieve the above very challenging design targets, research is focused on the following topics:
♦         Beam-channel radio characteristics: The path loss, shadowing and delay spread characteristics over a beam channel are the key parameters that impact the design of the new RATs with a beam-domain processed LSA array. The measurement and modeling of the beam-channel radio characteristics is the first step in the system design and performance evaluation. The measurement and modeling should be done for frequency bands above 6 GHz considering different beam-widths. The issues of spatial consistency and effects of specular scattering should be investigated in the research.
♦          Multiple access technology: Theoretically, the beam-domain processed LSA technology is workable with different types of multiple access including OFDMA/DFT-spread OFDMA, single-carrier multiple access, SIC-amenable multiple access (SAMA), and others. It is also workable with different waveform design such as FBMC (filter bank multi-carrier) and UFMC (universal-filtered multi-carrier). More importantly, it has the potential to be used in FDD systems because of its much smaller number of beam-channels seen by UEs. A multiple access technology for the new RATs should include the designs of system numerology, waveforms, TDD/FDD operation, downlink/uplink reference signals, and flexible access protocol, etc. 
♦          Network architecture: A two-layer network architecture consisting of a low-frequency macro cells that provide coverage and mobility management and high-frequency small cells that provide high data rates is a popular approach. On the other hand, the stand-along network architecture consists of merely small cells that provide high data rate access over a wide area should also be investigated. The C-RAN architecture with flexible backhauling and front-hauling and techniques for tight beam-channel cooperation between small cells should be investigated.  
♦          Beam-forming technology: Accurate beam-forming is an essential step in applying beam-domain processed LSA technology to increase data rate and system capacity. A configurable and scalable LSA architecture with a varied number of RF transceiver units provide a flexible way of using the LSA array in a cost effect way. The calibration and compensation of RF effects such as antenna mutual coupling, and I/Q and transceiver units imbalances in the LSA array is a key issue to be addressed for having accurate beam-forming.
 ♦         Beam-domain signal processing: In a RAT with a beam-domain processed LSA array, the key beam-domain processing techniques include beam-book design that takes RF effects into consideration, users’ beam finding and tracking, beam-domain SU/MU-MIMO operations. Low-complexity techniques are essential to lower the system overall complexity.


cron web_use_log