Date post: | 13-Apr-2017 |
Category: |
Technology |
Upload: | edouard-deberdt |
View: | 58 times |
Download: | 0 times |
Technology for a better society
Perspectives and challenges with millimetre-wave
(mmW) communications
TEK.NO konferansen 28-29/05/2015 – Kragerø
Isabelle Tardy & Jan E. Håkegård
Technology for a better society
Answering 5G soaring capacity need
• Three paths to increase the system
capacity
1. Allocate additional spectrum through
re-farming or introducing new bands,
2. Improve the spectral efficiency of the
technologies (e.g., MIMO),
3. Cell densification.
• But, decreasing cell size may induce
more interference, affecting in turn the
spectrum efficiency.
Non-interfering frequencies within dense
deployments, e.g., mmW.
2
(from I.Wang et al. Qualcomm tech., "A holistic view on hyper-dense
heterogeneous and small cell networks", IEEE Comm. Mag, June 2013)
Technology for a better society
Which mm-bands? How much gain is expected?
Frequency
band (GHz)
Available
bandwidth
(GHz)
Duplexing Max EIRP (dBm) Comments
28 1.3 Paired bands
with center gap
85
Freed FS spectrum (ITU2).
P-P, P-MP, M-P-P for feeding small cells and large
cells.
32 1.6 Paired bands
with center gap
80
Worldwide freed FS spectrum.
P-P, P-MP, M-P-P for feeding small cells and large
cells.
40 3 Duplex-
neutral blocks
80 Freed FS spectrum (ITU1).
P-P, P-MP, M-P-P for feeding small cells and large
cells. Block assignment recommended.
60
"V-band"
7 Duplex-
neutral blocks
Up to 85 P − MPUp to 55 P− P
Depending on
antenna gain (EU).
Worldwide. HDFS in 57-59 and 64-66 GHz. Unlicensed
for SRD with EIRP=20 dBm.
Gmin=30 dBi (EU only, [1]).
High attenuation possible, small range or indoor
usage preferred.
70&80
"E-band"
10 Duplex-
neutral blocks
Up to 85
Depending on
antenna gain (EU).
Worldwide FS/HDFS in 71-76 and 81-86 GHz bands,
but national allocations vary.
Gmin=38 dBi, Pmax=10dBm (EU only, [1]).
Small range (but higher than at 60 GHz, see next §).
P-P, P-MP, M-P-P for feeding small cells and large
cells.
3
mmW = [30 -300] GHz, per def.
Internationally agreed, license-exempt and lack sharing constraints
Technology for a better society
mmW usages
• Historically, mmW communications for
point-to-point (P-P) terrestrial and
satellite radio links.
• More recently, solutions for indoors short-
distance, high-capacity networks
• Outdoor backhauling to cellular base
stations or between them (x2-interface).
– mmW are not expected to be backward-
compatible with cmW radio access technologies
– mmW are not expected to be standalone
– Control information through the macrocells
(fallback and directionality tradeoffs)
4
(from I. Tardy and J.E. Håkegård, "Millimeter-wave communication in 5G"
chap.5 of the white paper of the IEEE Special Interest Group on Cognitive
Radio in 5G, Novel Spectrum Usage Paradigms for 5G. Editors M. Mueck,
W. Jiang, G. Sun, H. Cao. E. Dutkiewicz, S. Choi, Nov. 2014.)
Technology for a better society
Propagation characteristics in mmW range
• Indoor– A number of different semi-empirical models exist, with
clustering of signal energy both in the temporal and spatial domains
– Offices/cubicles, hospitals, etc…
• Outdoor– Direct measurements and a few modelling examples, e.g., in
New York city or on university campus
– Lack of good and exhaustive channel models
• Free space path attenuation increases with the square of frequencies, but OK for short distances
• Atmospheric losses and rain attenuation, but OK for short distances
• Blocking causes outage
5
Technology for a better society
mmW communication standards – 60 GHz
• Mosty indoor in the 57-66 GHz range, e.g., as replacement for High-Definition Multimedia Interface (HDMI) cables or similar.
• IEEE 802.11 ad (WiGig): • Adaptive beamforming, Single Carrier and optional
OFDM, backward compatibility with 2.4 and 5 GHz IEEE 802.11 n/ac
• Rates up to 7 Gbit/s (x10 highest 802.11n rate) and channel BW=2.16 GHz
6
(from Wilocity )(from Tensorcom)
(tri-band routers
from Netgear)
Technology for a better society
mmW MIMO (1/2)
• Small form factor
• Beamforming to compensate for
losses
• Should strike a balance between
performance, complexity and power
– Analogue beamforming – simple
– Digital beamforming – better
performance, but costly DAC
• Design: number of antennas
addressing narrow beams, energy
efficiency, spectrum efficiency…
• Spatial multiplexing for additional
gain, capturing energy in selected
directions
7
2-stage, hybrid
BS with several antennas
Technology for a better society
mmW MIMO (2/2)
• Adaptive beamforming requires
precise channel state information
(CSI)
• But considering a spare channel
matrix, the Angle of Arrival/Departure
(AoA/AoD) are expected sufficient
• Hybrid beamforming is interesting in
the angular vicinity of the analog
counterpart
8
2K=subcarriers, N=transceivers, D=precision of AoD
Hybrid Analog
Technology for a better society
Beam-searching and transmission scheduling in mmW
• MAC design in mmW will differ from mW
1. Control channel architecture
2. Initial access, handover
3. Resource allocation
9
1
2 3
time
Technology for a better society
1. Four design options of the control channel
1. Omni-directional mmW for broadcasting inside
small cells
2. Semi-directional mmW also for broadcasting
inside small cells. Additionally, the channel can be
used for feedback such as Hybrid Automatic
Repeat reQuest (HARQ)
3. Fully-directional mmW if a good alignment
between BS and UE can be realized. HARQ is
indeed possible in this option.
4. Omnidirectional mW as a fallback option
10
Insufficient information to extract spatial
synchronization due to different
propagation characteristics
Control channel (omni) with a more limited
range than the data channel (directional)
Technology for a better society
2. Initial access and handover strategies
• Beam-searching: balance between narrower beams with better directivity gain
and wider beams with faster beam-search.
– Wider beams in the search phase, narrower beams for scheduling traffic
• Given a discovery probability, the full directionality option (3) requires more
time than a semi-directionality one (2).
• A priori angular selection to reduce the spatial search overhead while
providing a minimum level of coverage will lead to a good compromise.
• Additionally, exploit the channel or beam reciprocity of TDD channels to
determine the best beam that the terminal must use.
– Channel reciprocity holds if the duplexing time is much shorter than the coherence time of
the channel. mmW channels have a coherence time an order of magnitude lower than mW, as
the Doppler shift scales linearly with frequency.
mmW channels are best suited to scenarios with low mobility.
• More blockage, more handovers. Several active beams from UE to several BS?
11
Technology for a better society
3. Resource allocation in mmW
• Once the search is over, scheduling
can be done on a space diversity
basis, allocating a frequency slot for
each angular-separated beam
– time-frequency-space (t, f, s) allocation
• An extra relay station can be useful to
counter blockage.
12
Semi-directional Fully directional
(from H. Shokri-Ghadikolaei et al., "Millimeter wave cellular networks: a MAC
layer perspective", arXiv:1503.00697, mars 2015)
Technology for a better society
• Develop a flexible SW&HW mmW platform
• Develop competence in model based DSP design using Xilinx System Generator
and MATLAB & Simulink
• Demonstrate the platform in one or several use cases
• Facilitate contribution in national and international R&D projects within mmW and
5G
13
60 GHz communication project
Technology for a better society 14
HW&SW platform
Xilinx ZC706 evaluation board
Zynq7000
PC
MATLAB/Simulink
Xilinx System Generator
A/C and D/C
60 GHz
RX/TX
RF front-end
60 GHz
RX/TX
RF front-end
Model-Based DSP Design
• Contains two 2x2 MIMO transceivers
• Frequency range 57-64 GHz
• EIRP=23.5 dBm
• Four-channel, 16-bit A/D up to 370 MSPS
• Four-channel, 16-bit D/A up to 2.5 GSPS
Technology for a better society
• Channel measurements at 60 GHz
• Indoor and outdoor
• Path loss
• Fading characteristics (impulse response, Doppler spread)
• Channel modelling
• Testing and demonstration of
• Communication protocols (PHY and MAC layers)
• 2x2 MIMO techniques
• Beam-forming (incl. DoA estimation)
• Spatial multiplexing & diversity gain
16
Potential use cases