Interference Cancellation as a Mobile Enhancement to Improve Spectral Efficiency

April 2006

Louis Scharf, Chief Scientist

2

Narrative Outline

The Shannon Game is to trade power efficiency for spectral efficiency in the wireless network. To score in the game we need to match transmitters and receivers to the channel effects (interference). This impels us to understand:

• Interference: Its sources and its effects

• Interference Cancellers

Then we can analyze:

• Performance of Interference Cancellers

• Economic Impact of Interference CancellersTh

roug

hput

(kbp

s)

Power (watts)

Shannon limit

Lower power at no decrease in data rate

Higher data rate at no increase in power

3

Interference is the Key Limiter ofCode-Based Network Performance

•As reported by Morgan Stanley, Feb. 1, 2005. 3G Test conducted in London.

Data rate increase

Dramatic demonstration of the Shannon game:• For increased data rates, only users near cell center have sufficient power, thus reducing effective cell

size. So increase power. But this increases interference, producing no useful effect for power, unless an interference cancellation receiver is matched to channel.

Data rate and capacity reduction; Quality of service degradation• Degraded reception leads to lower data transfer rates• Majority of users effectively at edge of cell• Interference causes base stations to broadcast more power to each terminal• The number of terminals that can be supported simultaneously is reduced• Increase in coverage holes leads to more frequent dropped calls

4

Evolution of Interference Cancellation Receivers from CDMA2000 to 1xEV-DO and WCDMA/HSDPA

Historically, interference cancellation for all channels has been considered too complex and costly to deploy. Key breakthrough is an interference cancellation solution that can be implementedwithin a baseband chip.

There has been a belief that a PN-averaged equalizer bounds the performance of an advanced receiver for 1xEV-DO. Recent developments prove that this equalizer bound can be breeched with interference cancellation.

Interference cancellation integrates with receive diversity and extends to MIMO-OFDM.

Interference cancellation for WCDMA/HSDPA is a study item in 3GPP RAN4. TensorComm is a participant and contributor.

5

TensorComm’s technology roadmap continues to improve gains of Interference Cancellation Technology (ICT™)

March: Operator proves substantial capital savings in network capacity and performanceApril: Publicly announce ICTJuly: Extend solution to WCDMA.November: Announce next generation ICT for 1xEV-DO and HSDPA

2004 2005 2006 Future …..

Aug 04: System simulations prove viability of ICTNov 04: ICT gains demonstrated in field trials on commercial CDMA2000 network using integrated FPGA platforms

Jan: ICT 440 data throughput increases of 50%+ proven in system simulationsGains up to 6 dbFeb: Results submitted to 3GPPRequested by operators to conduct network trials

•Improvements to dB gain; size reduction•S-DMB applications•Base station solution•MIMO-OFDM solution

ICT 220 ICT 440

TensorComm continues to fortify its Intellectual Property and toconduct evaluations in cooperation with operators and chipset vendors

6

Characterizing Interference

The problem is particularly problematic when coded sub-channels are densely populated and/or the mobile terminal finds itself in an edge of cell condition. The technology challenge is to cancel these effects, at manageable complexity, in order to maintain cell size and/or increase data rates.

Channel effects defeat all attempts to maintain the orthogonality between coded sub-channels in a multi-access system. As a consequence there is interference between sub-channels, and it comes in two varieties:

• intra-cell interference due to multi-path structure• inter-cell interference due to interfering base stations

Ior (1) Ior (2)

Ec Ec

F1F2

Ior (1) Ior (2)

Ec Ec

F1F2

7

A Representation of Advanced Receivers

8

Characteristics of Advanced Receivers

Only ICT matches to inter-cell and intra-cell interference, at the symbol rate.ICT removes interference from all channels (not just pilot).ICT fills performance gaps that arise between RAKEs, equalizers and MUDs. ICT subsumes equalizers, and it fits within the architecture of a RAKE receiver, with no impact on RF design.ICT integrates with MRD.

RAKE Equalizer ICT

Matches to intra-cell multi-path

√ √ √

Matches to inter-cell multi-path

√(in soft hand-off) √

Manages intra-cell code correlation

√(@ doppler rate)

√(@ symbol rate)

Manages inter-cell code correlation

√(@ doppler rate)

√(@ symbol rate)

9

ICT Fills Performance Gaps

Brackets represent the performance gaps that ICT fills by riding as close to the MUD

line as possible, at manageable hardware

complexity

Figure not scaled

10

HSET_6 QPSKMid Geometry (Ior/Ioc = 5 dB)2 Basestations, 1 InterferingEc/Ior = -6 dB for HSDPA (10-code)

% Increase in User

Throughput

Current Receiver (RAKE) * 433 595 270%

Type 2 Future Receiver (LMMSE equalizer) 976 988 90%

Type 1 Dual Antenna Receiver (RD) 1181 1220 55%

Type 3 (RD + LMMSE equalizer) 1651 1674

TensorComm ICT ** 1825 1899

Pedestrian BVehicular 30km/hr

Average ICT Gain Over

User Throughput (kbps)

ICT Increases User Data Throughput

* Rake by itself will not provide meaningful data throughputs. For many channel conditions and geometries, advanced receivers will be necessary for any realistic applications beyond voice.

** Interference is modeled as a true node B, per 3GPP 25-101

TensorComm has completed extensive link-level simulations conducted in conjunction with multiple WCDMA chipset vendors.These comparisons are conservative because RAKE, equalizer and RD use AWGN model for interfering base stations.A sample of similar results included in contribution to 3GPP : RAN4 document number R4-060043.

11

1.9dB

ICT Optimized for 1xEV-DO Data Networks

Live data collects from U.S. 1xEV-DO markets indicate opportunity for ICT gain over RAKEs and equalizers in more than 60% of network conditions.When gains are averaged over observed network conditions, the sector throughput gains are increased by more than 40%.

Sample Test: Two-cell; no multi-path, Veh A

12

ICT Gains for CDMA2000 1X

• Test cases are representative of interference-limited network scenarios derived from field data for a near capacity network :

•Moderate to heavy loading•Cell edge•Multi-path and fading•Multiple sectors

Channel Condition DefinitionType I Stationary; non-fadingType II 30 km/hr; Path profile=[0,-3]Type III 100 km/hr; Path profile=[0,-3]

A Cell edge, 44 of 64 users

2 / 2 Type I 4.7 dB

B Cell edge, 64 of 64 users

2 / 2 Type I 2.4 dB

C Mid cell, 44 of 64 users,

unequal pilots (0, -6dB)

2 / 2 Type I 4.8 dB

F Cell edge, 44 of 64 users

4 / 2 Type III 2.1 dB

Test Case Test Environment Description

Interfering Number of Paths / Sectors

Channel Condition

ICT Gain (dB)

D Cell edge, 44 of 64 users

4 / 2 Type II 3.5 dB

E Cell edge, 3 way softhandoff,

44 of 64 users

6 / 3 Type II 2.0 dB

13

F o rw a r d T r a f f ic C h a n n e l P o w e r v s . T im e7 5 % O C N S

1 5

1 8

2 1

2 4

2 7

16:1

1:53

.560

16:1

2:09

.440

16:1

2:25

.320

16:1

2:41

.240

16:1

2:57

.120

16:1

3:13

.000

16:1

3:28

.920

16:1

3:44

.800

16:1

4:00

.680

16:1

4:16

.560

16:1

4:32

.480

16:1

4:48

.360

16:1

5:04

.240

16:1

5:20

.160

16:1

5:56

.400

16:1

6:12

.320

16:1

6:28

.200

16:1

6:44

.079

16:1

6:59

.960

16:1

7:15

.880

16:1

7:31

.760

16:1

7:47

.640

16:1

8:43

.060

16:1

8:58

.980

16:1

9:14

.860

16:1

9:30

.740

16:1

9:46

.620

16:2

0:02

.540

16:2

0:18

.420

16:2

0:34

.300

16:2

0:50

.220

16:2

1:06

.100

16:2

1:21

.980

16:2

1:37

.860

16:2

1:53

.780

16:2

2:09

.660

16:2

2:25

.540

16:2

2:41

.460

16:2

2:57

.340

C a n c e lle r O f f C a n c e lle r O n C a n c e lle r O f f C a n c e lle r O n

Fwd

TCH

Pow

er (d

bm)

ICT Trials Conducted in a Major Operator’s Controlled Laboratory with Live Base Stations

• Operator’s controlled laboratory configuration consisted of a two-sector, live base station and fading simulator.

• ICT 220 prototype manually switched between ICT220-enabled and disabled modes during the call.

• Logs capture the base station power transmitted during the call.

ICT-enabled devices draw less power from the base station

14

1 9

2 0

2 1

2 2

2 3

2 4

2 5

2 6

2 7

2 8

2 9

3 0

3 1

3 2

3 3

3 4

3 5

3 6

15:4

7:46

.5

15:4

7:51

.0

15:4

7:55

.5

15:4

7:59

.7

15:4

8:04

.0

15:4

8:08

.3

15:4

8:12

.5

15:4

8:16

.8

15:4

8:21

.0

15:4

8:25

.8

15:4

8:30

.4

15:4

8:34

.8

15:4

8:39

.1

15:4

8:43

.5

15:4

8:48

.1

15:4

8:52

.5

15:4

8:56

.8

15:4

9:01

.1

15:4

9:05

.5

15:4

9:09

.7

15:4

9:14

.0

15:4

9:18

.4

15:4

9:22

.6

15:4

9:27

.1

15:4

9:31

.3

15:4

9:35

.6

15:4

9:40

.0

15:4

9:48

.4

15:4

9:53

.3

15:4

9:57

.8

15:5

0:02

.1

15:5

0:06

.5

15:5

0:10

.8

15:5

0:15

.3

15:5

0:19

.5

15:5

0:23

.8

T im e

Fwd

TCH

Pw

r (db

m)

ICT Proven on a Commercial CDMA2000 NetworkProto 1 -ICT onProto 2 -ICT off50 per Mov. Avg (Proto1)50 per Mov. Avg (Proto2)

• Low to moderate speed (3-30km/hr)

• Base station logs

Evaluated on a major operator’s commercial network on a handset designed by one of the world’s top manufacturers. Base station power reduction when the network needs it the most, with edge case gains, particularly in blocking situations, of up to 6dB.Over 60% gain in network capacity.

15

Operators Can Derive Dramatic Economic Benefits with ICT

ICT gain translates into delay of network infrastructure

• Each incremental ICT-enabled terminal provides a benefit to the consumer and operator.• Provides no less than $30 benefit per handset• Additional spectrum expansion delayed or eliminated

0

100

200

300

400

500

600

Q0

Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

Q9

Q10

Q11

Q12

Q13

Q14

Q15

Q16

Q17

Q18

Q19

Q20

Q21

Q22

Q23

Q24

Q25

Q26

Q27

Q28

Sup

port

ed D

ata

User

s P

er S

ecto

r

Data User Demand

Sector Capacity (no ICT)

Sector Capacity (ICT)

Sector Capacity w ith CarrierAdd or Cell Split (no ICT)Sector Capacity w ith CarrierAdd or Cell Split (ICT)

Data user Demand @ 20% Annual Growth Rate

Initial state: 55% load

Breakpoint for non-ICT network at Q10

Breakpoint for ICT network at Q21

Additional 103 users/sector and 11 quarter delay in infrastructure deployment

66% increase in users: delta represents 206 additional users/sector with ICT network

*ICT penetration at 5% of all new handsets for first year, increasing to 25%, 50%, 75% and 100% each following year

Second breakpoint for non-ICT network at Q25

Second Carrier Sector Capacity with ICT

Second Carrier Sector Capacity with non-ICT

0

100

200

300

400

500

600

Q0

Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

Q9

Q10

Q11

Q12

Q13

Q14

Q15

Q16

Q17

Q18

Q19

Q20

Q21

Q22

Q23

Q24

Q25

Q26

Q27

Q28

Sup

port

ed D

ata

User

s P

er S

ecto

r

Data User Demand

Sector Capacity (no ICT)

Sector Capacity (ICT)

Sector Capacity w ith CarrierAdd or Cell Split (no ICT)Sector Capacity w ith CarrierAdd or Cell Split (ICT)

Data user Demand @ 20% Annual Growth Rate

Initial state: 55% load

Breakpoint for non-ICT network at Q10

Breakpoint for ICT network at Q21

Additional 103 users/sector and 11 quarter delay in infrastructure deployment

66% increase in users: delta represents 206 additional users/sector with ICT network

*ICT penetration at 5% of all new handsets for first year, increasing to 25%, 50%, 75% and 100% each following year

Second breakpoint for non-ICT network at Q25

Second Carrier Sector Capacity with ICT

Second Carrier Sector Capacity with non-ICT

16

ICT Derives Significant Gains for 3G Data Networks

Code-based systems are interference limited.

Gains from interference cancellation are variable with network and channel conditions… in realistic channel conditions, up to 6 dB and an increase in user data throughput no less than 50%.

• Higher data throughput to user • Higher network capacity• Reduced base station power (and interference)• Fewer dropped calls in heavily loaded networks• Wider coverage area

17

ICT Derives Significant Gains for 3G Data Networks

ICT outperforms equalizers, is comparable to MRD, and approximates MRD plus equalizers, at a fraction of the cost.

ICT is a hardware efficient design, ready for integration and optimized for data networks. ICT has been proven in simulations and commercial drive tests.

Technology roadmap for ICT extends beyond 3G technologies to MIMO-OFDM.