How many antennas does it take to get wireless access? -The story of MIMO n Benjamin Friedlander n Department of Electrical Engineering n University of.

How many antennas does it take to get wireless access?-The story of MIMO

Benjamin Friedlander Department of Electrical Engineering University of California at Santa Cruz

Phone: 831-459-5838 [email protected]

April 25, 2005

What is MIMO?

MIMO Multiple Input Multiple Output Using multiple antennas on both sides

of a communication link SISO

Single Input Single Output SIMO

Single Input Multiple Output

So what do we so with multiple antennas?

Phased Array / Beamformer

Beam Pattern & Gain

Array Single Element

Array gain = maximum power density relative to omni-directional antenna

Space Division Multiple Access

Traditional wireless resources: frequency and time

New resource: space

Large capacity gains possible (in theory)

User #1

Basestation

User #2

User #3

And then there was MIMO …

SDMA - Double the capacity

MIMO - Double the capacity?

Scattering & Multipath

TX RX

MIMO: Spatial Multiplexing

TX RX

1/2

1/2

P/2

P/2

2 2

2 2

P

P

SISO

TX RX

1/2

1/2

P P

1

1

1

1

MIMO: Beamforming*

TX RX

1/2

1/2

P/2

P/2

2 2

2 2

P

P

2PP

* Non standard use of term

M x M System

Spatial multiplexing – M channels with gains depending on channel. Average SNR same as SISO.

Beamforming - single channel with SNR gain relative to SISO.

Various intermediate combinations possible

Combination of Multiplexing and Beamforming

TX RX

SIMO

TX

RX

1/2

1/2

P

1 2

1 2

2P

P

P

MIMO Performance

Depends on the channel gains Assuming channel gains random,

independent: MIMO capacity approximately M times SISO capacity due to spatial multiplexing

Theoretical Capacity

-20 -15 -10 -5 0 5 10 15 20 250

5

10

15

20

25

30

SNR [dB]

Capacity (M=4)

siso

simomimo

Bits/sec/Hz

Large Angular Spread

TX RX

Small Angular Spread

TX RX

Theoretical Capacity

5 10 15 20 25 30 35 40 45 500

5

10

15

20

angle spread [dB]

capacity, snr = 20 dB, M = 4

sisosimomimo

Bits/sec/Hz

Conclusion #1 MIMO is best when SNR and angular

spread are large Small angular spread, or presence of a

a dominant path (e.g. LOS) reduce MIMO performance

Question: what percentage of cases are “MIMO friendly”?

Beamforming – SNR GAIN

Multiple antennas can be used to provide increased SNR

SNR gain has two components Array gain – increasing the average

power Diversity gain – decreasing power

fluctuations and thereby decreasing required margin

Conclusion #2 Consider a system with a fixed modulation –

say 64-QAM. Spatial multiplexing: increases throughput, not

range* Beamforming: increases range (SNR), not

throughput Possible to do combinations of multiplexing

and beamforming Additional range/throughput tradeoff using

variable modulation

* Ignoring coding effects

The Promise of MIMO

Increased throughput without requiring more spectrum

Increased range without requiring more transmit power

Word of Caution

Smart antennas & MIMO can provide large performance gains in theory

In practice implementation issues and system issues often erode much of these gains

Some of the issues

What are we comparing to? Switched diversity SIMO (RAKE receiver)

Channel Estimation Performance of multi-user system

dominated by worst user (low SNR, small angle spread)

Status of MIMO

802.11n Pre-n products 802.16 3G & beyond

Some 802.11n Proposed Specs

TGn Sync 2x2, 20 MHz – 140 MBPS 4x4, 40 MHz – 630 MBPS

WWiSE 2x2, 20 MHz – 135 MBPS 4x4, 40 MHz – 540 MBPS

So what is the real MIMO advantage? Most performance claims published so

far are not well documented and impossible to evaluate

Need testing over a broad range of deployments and operating conditions, in carefully designed experiments

Only time will tell …

Final Words Many antennas are better than one Standardization and reduced costs are making

MIMO a viable technology Current MIMO systems – impressive achievement MIMO improves performance, but:

Your performance may vary … Thorough performance evaluation not yet

available Differences likely between expectations and

reality

For additional information

Please contact [email protected] Related talks:

Wireless Facts and Fiction Multi-access methods: TDMA, FDMA, CDMA, OFDMA

– so what comes next? Wireless in the wild west: operating in the unlicensed

spectrum. Communicating on the move – mobility and its

limitations The amazing story of ultra-wideband