Friday, April 15, 2016

Facebook Project ARIES Uses Massive MIMO to Radically Improve Specrum Efficiency

In addition to pioneering a new fixed wireless access system for Internet access, Facebook also is working on a new spectrally-efficient air interface based on use of massive MIMO antenna arrays.

Project ARIES, presently in test mode, features a base station with 96 antennas, supporting 24 simultaneous streams.

Facebook so far has achieved 71 bps/Hz, and Facebook engineers are aiming for supplying 100 bps/Hz eventually.

By way of comparison, that would surpass the efficiency of Long Term Evolution and vastly outperform most other interfaces ever commercially deployed.

In large part, that effort is based on use of massive MIMO.

Service
Standard
per carrier
per one spatial stream
(Mbit/s)
per carrier
(MHz)
System spectral efficiency (R/B)/K
((bit/s)/Hz per site)

1G cellular
NMT 450 modem
0.0012
0.025
0.064
1G cellular
AMPS modem
0.0003[2]
0.030
0.0015
2G cellular
0.013 × 8 timeslots = 0.104
0.2
0.17[4] (in 1999)
2G cellular
0.013 × 3 timeslots = 0.039
0.030
0.45[4] (in 1999)
2.75Gcellular
CDMA2000 1× voice
0.0096 per phone call × 22 calls
1.2288
0.172 (fully loaded)
2.75Gcellular
0.384 (typ. 0.20)
0.2
0.33[4]
2.75Gcellular
0.384 (typ. 0.27)
0.200
0.45[4]
3G cellular
WCDMA FDD
0.384
5
0.51
3G cellular
CDMA2000 1x PD
0.153
1.2288
0.1720 (fully loaded)
3G cellular
CDMA2000 1×EV-DO Rev.A
3.072
1.2288
1.3
FixedWiMAX
96
20
1.2
3.5G cellular
21.1
5
4.22
4G MBWA
3.9
0.625
7.23
4G cellular
81.6
20
16.32
4G cellular
75
20
30
54
20
0.900
72.2 (short GI)
20
1.2
433.3 (short GI)
80

6756
2160
1
4 timeslots = 0.036
0.025

0.576 to 1.152
1.712
0.08 to 0.17
DAB with SFN
0.576 to 1.152
1.712
0.34 to 0.67
31.67 (typ. 22.0)
8
0.55
DVB-T with SFN
31.67 (typ. 22.0)
8
4.0 (typ. 2.8)
5.5 to 11
8
0.14 to 0.28
DVB-H with SFN
5.5 to 11
8
0.68 to 1.4
38
6
N/A
ADSL2 downlink
12
0.962
N/A
ADSL2+ downlink
28
2.109
N/A
V.92 downlink
0.056
0.004
N/A


No comments:

Post a Comment

Is Sora an "iPhone Moment?"

Sora is OpenAI’s new cutting-edge and possibly disruptive AI model that can generate realistic videos based on textual descriptions.  Perhap...