practical, real-time, full-duplex...

Practical, Real-time,Full-Duplex Wireless

Mayank Jain, Jung Il Choi, Taemin Kim,Dinesh Bharadia, Kannan Srinivasan, Siddharth Seth,

Philip Levis, Sachin Katti, Prasun Sinha

September 22, 2011

1

Single channel full-duplex

2

TX

RX

RX

TXNode 1 Node 2

Mobicom’10[1]:

Antenna Cancellation + other techniques

3

The story so far...

d d + λ/2

TX1 TX2RX

[1] Choi et al. “Achieving single channel, full duplex wireless communication”, Mobicom 2010

Mobicom’10[1]:

Antenna Cancellation + other techniques

4

The story so far...

d d + λ/2

TX1 TX2RX

• Frequency dependent, narrowband

• Requires manual tuning

• Two transmit antennas cause complex far-field behavior

• New full-duplex radio design: signal inversion cancellation

• Wideband, frequency independent

• Adaptive

• One transmit antenna design

• Real-time full-duplex MAC layer implementation

• Show MAC layer gains with full-duplex

5

Contributions

• RF Design using Signal Inversion

• Adaptive RF Cancellation

• System Performance

• Implications to Wireless Networks

• Open Questions

Talk Outline

6

• RF Design using Signal Inversion

• Adaptive RF Cancellation

• System Performance

• Implications to Wireless Networks

• Open Questions

Talk Outline

7

The challenge of full-duplex

➔ Very strong self-interference: ~70dB for 802.11

8

TX

RX

RX

TXNode 1 Node 2

Self-Interference

9

TX

RX

RX

TXNode 1 Node 2

Self-Interference

Combine RF and digital techniques for cancellation

The challenge of full-duplex

➔ Very strong self-interference: ~70dB for 802.11

Cancellation using Phase OffsetSelf-Interference

Cancellation Signal

∑

Cancellation using Phase Offset

11

Self-Interference

Cancellation Signal

∑

Self-Interference

Cancellation Signal

∑

Frequency dependent, narrowband

Cancellation using Signal Inversion

12

Self-Interference

Cancellation Signal

∑

Cancellation using Signal Inversion

13

Self-Interference

Cancellation Signal

∑

Self-Interference

Cancellation Signal

∑

Frequency and bandwidth independent

Time

14

Xt +Xt/2

-Xt/2

BALUN

BALUN : Balanced to Unbalanced Conversion

Time

15

TX RX

TXRF Frontend

Xt

+Xt/2 -Xt/2

∑

RXRF Frontend

Xt +Xt/2

-Xt/2

BALUN

BALUN : Balanced to Unbalanced Conversion

Cancellation Signal

Time

16

TX RX

TXRF Frontend

Xt

+Xt/2 -Xt/2

∑

RXRF Frontend

Xt +Xt/2

-Xt/2

BALUN

Over the air attenuation and delay

Time

17

TX RX

TXRF Frontend

Attenuator andDelay Line

Xt

+Xt/2 -Xt/2

∑

RXRF Frontend

Xt +Xt/2

-Xt/2

BALUN

Signal Inversion Cancellation

Over the air attenuation and delay

+Xt/2

• Measure wideband cancellation

• Wired experiments

• 240MHz chirp at 2.4GHz to measure response

Time

Signal Inversion Cancellation: Wideband Evaluation

18

TX

RX

Signal Inversion Cancellation Setup

∑ TX RX

Phase Offset Cancellation Setup

∑RF

Signal Splitter

Xt +Xt/2

-Xt/2

Xt

+Xt/2

λ/2 Delay

Time

19

Lower isbetter

Higher isbetter

Time

20

~50dB cancellation at 20MHz bandwidth with balun vs ~38dB with phase offset cancellation.

Significant improvement in wideband cancellation

Lower isbetter

Higher isbetter

Time

21

• From 3 antennas per node to 2 antennas

• Parameters adjustable with changing conditions

Attenuator andDelay Line

TX RX

TXRF Frontend

Xt

+Xt/2 -Xt/2

∑

RXRF Frontend

Other advantages

• RF Design using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation

• System Performance

• Implications to Wireless Networks

• Open Questions

Talk Outline

22

• Need to match self-interference power and delay

• Can’t use digital samples: Saturated ADC

Adaptive RF Cancellation

23

TX RX

Attenuation & Delay

Wireless Receiver

Wireless Transmitter

RF Cancellation

TX Signal Path RX Signal Path

RF ReferenceΣ

Balun

• Need to match self-interference power and delay

• Can’t use digital samples: Saturated ADC

Adaptive RF Cancellation

24

RSSI : Received Signal Strength Indicator

TX RX

Attenuation & Delay

Wireless Receiver

Wireless Transmitter

RF Cancellation

TX Signal Path RX Signal Path

RF ReferenceΣ

Balun

RSSI

• Need to match self-interference power and delay

• Can’t use digital samples: Saturated ADC

Adaptive RF Cancellation

25

Use RSSI as an indicator of self-interference

TX RX

Attenuation & Delay

Wireless Receiver

Wireless Transmitter

RF Cancellation

TX Signal Path RX Signal Path

RF ReferenceΣ

Balun

RSSI

Control Feedback

26

Objective: Minimize received power

Control variables: Delay and Attenuation

TX RX

Attenuation & Delay

Wireless Receiver

Wireless Transmitter

RF Cancellation

TX Signal Path RX Signal Path

RF ReferenceΣ

Balun

RSSI

Control Feedback

27

Objective: Minimize received power

Control variables: Delay and Attenuation

28➔ Simple gradient descent approach to optimize

Objective: Minimize received power

Control variables: Delay and Attenuation

29

Off-the-shelf electronically tunable hardware approximation: QHx220 noise canceler

Gain Q

Gain I

λ/4 Delay

Σ Σ

InterferenceSample

Signal + Interference

CancellationSignal Clean

Signal

30

Off-the-shelf electronically tunable hardware approximation: QHx220 noise canceler

Gain Q

Gain I

λ/4 Delay

Σ Σ

TX RX

Wireless Receiver

Wireless Transmitter

TX Signal Path RX Signal Path

RSSI

Control FeedbackBalun

31

Off-the-shelf electronically tunable hardware approximation: QHx220 noise canceler

32

Typical convergence within 8-15 iterations (~1ms total)

• RF Design using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation: ~1ms convergence

• System Performance

• Implications to Wireless Networks

• Open Questions

Talk Outline

33

Digital Cancellation

34

• Measure residual self-interference after RF cancellation

• Subtract self-interference from received digital signal

35

Digital Interference Cancellation

TX RX

Attenuation & Delay

RF ➔ Baseband

ADC

Baseband ➔ RF

DAC

Encoder Decoder

Digital Interference Reference

RF Cancellation

TX Signal Path RX Signal Path

RF ReferenceΣ

FIR filter ∆

RSSI

Control Feedback

Channel Estimate

Balun

Bringing It All Together

36

Channel Coherence

~3dB reduction in cancellation in 1-2 seconds

~6dB reduction in <10 seconds

37

Performance

• WiFi full-duplex: with reasonable antenna separation

• Not enough for cellular full-duplex: need 20dB more

• RF Design using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation: ~1ms convergence

• System Performance: ~73dB cancellation

• Implications to Wireless Networks

• Open Questions

Talk Outline

38

• Breaks a basic assumption in wireless

• Can solve some fundamental problems with wireless networks today[1,2]

• Hidden terminals

• Network congestion and WLAN fairness

Implications to Wireless Networks

39

[1] Choi et al. “Achieving single channel, full duplex wireless communication”, in Mobicom 2010[2] Singh et al. “Efficient and Fair MAC for Wireless Networks with Self- interference Cancellation”, in WiOpt 2011

• WARPv2 boards with 2 radios

• OFDM reference code from Rice University

• 10MHz bandwidth OFDM signaling

• CSMA MAC on embedded processor

• Modified for full-duplex

Implementation

40

• CSMA/CA can’t solve this

• Schemes like RTS/CTS introduce significant overhead

APN1 N2

Current networks have hidden terminals

Mitigating Hidden Terminals

41

• CSMA/CA can’t solve this

• Schemes like RTS/CTS introduce significant overhead

APN1 N2

Since both sides transmit at the same time, no hidden terminals exist

Current networks have hidden terminals

Full Duplex solves hidden terminals APN1 N2

Mitigating Hidden Terminals

42

• CSMA/CA can’t solve this

• Schemes like RTS/CTS introduce significant overhead

APN1 N2

Since both sides transmit at the same time, no hidden terminals exist

Current networks have hidden terminals

Full Duplex solves hidden terminals APN1 N2

Mitigating Hidden Terminals

43Reduces hidden terminal losses by up to 88%

Network Congestion and WLAN Fairness

Without full-duplex:

• 1/n bandwidth for each node in network, including APDownlink Throughput = 1/n Uplink Throughput = (n-1)/n

44

Network Congestion and WLAN Fairness

Without full-duplex:

• 1/n bandwidth for each node in network, including APDownlink Throughput = 1/n Uplink Throughput = (n-1)/n

45

With full-duplex:

• AP sends and receives at the same timeDownlink Throughput = 1 Uplink Throughput = 1

Network Congestion and WLAN Fairness

46

1 AP with 4 stations without any hidden terminals

Throughput (Mbps)Throughput (Mbps)Fairness (JFI)

Upstream DownstreamFairness (JFI)

Half-Duplex 5.18 2.36 0.845

Full-Duplex 5.97 4.99 0.977

Full-duplex distributes its performance gain to improve fairness

• RF Design using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation: ~1ms convergence

• System Performance: ~73dB cancellation

• Implications to Wireless Networks: Collisions, Fairness

• Open Questions

Talk Outline

47

48

Improving Full-duplex

• Non-linear channel response

• Non-linear channel responseReduce distortion: feedforward amplifiers

49

Improving Full-duplex

50

• Non-linear channel responseReduce distortion: feedforward amplifiersCompensate: non-linear digital cancellation

Improving Full-duplex

51

TX Signal RX Signal

Circulator

• Non-linear channel responseReduce distortion: feedforward amplifiersCompensate: non-linear digital cancellation

• Single antenna solution: circulators

Improving Full-duplex

52

• Non-linear channel responseReduce distortion: feedforward amplifiersCompensate: non-linear digital cancellation

• Single antenna solution: circulators

• MIMO full-duplex

Improving Full-duplex

53

Access Point networks

Full-duplex Networking

54

Access Point networksCellular networks

Cell Basestation

Relay

Full-duplex Networking

55

Access Point networks

Multi-hop Networks

Cellular networks

Cell Basestation

Relay

Full-duplex Networking

56

Access Point networks

Multi-hop NetworksSecure Networks[1,2]

Cellular networks

Cell Basestation

Relay

Full-duplex Networking

[1] Gollakota et al. “They Can Hear Your Heartbeats: Non-Invasive Security for Implantable Medical Devices.”, in Sigcomm 2011.[2] Lee et al. “Secured Bilateral Rendezvous using Self-interference Cancellation in Wireless Networks”, in IFIP 2011.

57

Access Point networks

Multi-hop NetworksSecure Networks[1,2]

Cellular networks

Cell Basestation

Relay

Full-duplex Networking

[1] Gollakota et al. “They Can Hear Your Heartbeats: Non-Invasive Security for Implantable Medical Devices.”, in Sigcomm 2011.[2] Lee et al. “Secured Bilateral Rendezvous using Self-interference Cancellation in Wireless Networks”, in IFIP 2011.

?

Summary

• Design for real-time full-duplex wireless

• Makes full-duplex WiFi possible

• Still some way to go for full-duplex cellular

• Made practical using adaptive techniques

• Rethinking of wireless networks

• WiFi: hidden terminals and fairness

• Many more possibilities

58

59

Thank You

Questions?

Backup

60

• Other cancellation techniques

Digital estimation for RF cancellation[1]

61

TX RX

RF ➔ Baseband

ADC

Baseband ➔ RF

DAC

TX Signal RX Signal

Σ

Baseband ➔ RF

DAC

Cancellation Signal

[1] Duarte et al. “Full-Duplex Wireless Communications Using Off-The-Shelf Radios: Feasibility and First Results.”, in Asilomar 2010.

• RF Cancellation using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation: ~1ms convergence

• Adaptive Digital Cancellation

• System Performance

• Implications to Wireless Networks

• Looking Forward

Talk Outline

62

Digital Cancellation

63

• Create a precise “digital replica” of the self-interference signal using TX digital samples

• Subtract self-interference replica from received digital signal

Requires ADC not saturated: RF cancellation

64

OFDM processing

SignalBand

65

OFDM processing

Sub-bands

66

OFDM processing

ChannelDistortion

67

OFDM processing

ChannelDistortion

Equalization

OFDM processing

68

ADC

RX RF Mixer

CarrierFrequency

Carrier FrequencyOffset Correction

Packet DetectFFT Engine

Channel Estimation

EqualizationDemapping

ChannelDistortion

Equalization

Step 1: Estimation

69

Self-interference Sounding

Preamble TrainingSequence Self-interference

Estimate

FIR Filter

ADC

RX RF Mixer

CarrierFrequency

Carrier FrequencyOffset Correction

Packet DetectFFT Engine

Channel Estimation

EqualizationDemapping

IFFT

Estimation includes effect of RF cancellation

Step 2: Cancellation

70

ADC

RX RF Mixer

CarrierFrequency

Carrier FrequencyOffset Correction

Packet DetectFFT Engine

Channel Estimation

EqualizationDemapping

TX Signal

+-

Self-interference Estimate

FIR Filter

IFFT

CancellationSignal

Step 2: Cancellation

71

ADC

RX RF Mixer

CarrierFrequency

Carrier FrequencyOffset Correction

Packet DetectFFT Engine

Channel Estimation

EqualizationDemapping

TX Signal

+-

Self-interference Estimate

FIR Filter

IFFT

CancellationSignal

30dB Cancellation

• RF Cancellation using Signal Inversion: ~50dB for 20Mhz

• Adaptive RF Cancellation: ~1ms convergence

• Adaptive Digital Cancellation: ~30dB cancellation

• System Performance

• Implications to Wireless Networks

• Looking Forward

Talk Outline

72

Attenuator

Phase Offset Cancellation: Block Diagram

d d + �/2TX1 TX2RX

RXRF Frontend

Digital Processor

TXRF Frontend

Power Splitter

73

-60

-55

-50

-45

-40

-35

-30

-25

0 5 10 15 20 25

RSS

I (dB

m)

Position of Receive Antenna (cm)

TX1 TX2

Only TX1 Active

Phase Offset Cancellation: Performance

74

-60

-55

-50

-45

-40

-35

-30

-25

0 5 10 15 20 25

RSS

I (dB

m)

Position of Receive Antenna (cm)

TX1 TX2

Only TX2 Active

75

Only TX1 Active

Phase Offset Cancellation: Performance

-60

-55

-50

-45

-40

-35

-30

-25

0 5 10 15 20 25

RSS

I (dB

m)

Position of Receive Antenna (cm)

NullPosition

TX1 TX2

76

Only TX1 ActiveOnly TX2 Active

Both TX1 & TX2 Active

Phase Offset Cancellation: Performance

-60

-55

-50

-45

-40

-35

-30

-25

0 5 10 15 20 25

RSS

I (dB

m)

Position of Receive Antenna (cm)

NullPosition

TX1 TX2

77

~25-30dB

Only TX1 ActiveOnly TX2 Active

Both TX1 & TX2 Active

Phase Offset Cancellation: Performance

78

What about attenuation at intended receivers?Destructive interference can affect this signal too!

• Different transmit powers for two TX helps

Single Transmit Antenna Two Transmit Antennas

0 10 20 30-30 -20 -10

0

10

20

30

-30

-20

-10

x axis (meters)

y ax

is (

met

ers)

-52 dBm

-58 dBm

0 10 20 30-30 -20 -10

0

10

20

30

-30

-20

-10

x axis (meters)

y ax

is (

met

ers)

-52 dBm

Sensitivity of Phase Offset Cancellation

79

Amplitude Mismatch between TX1 and TX2

Placement Errorfor RX

dB

Can

cella

tion

(dB)

Can

cella

tion

(dB)

Error (mm)

Higher is better

Higher is better

Amplitude Mismatch between TX1 and TX2

Placement Errorfor RX

dB

Can

cella

tion

(dB)

Can

cella

tion

(dB)

Error (mm)

80

30dB cancellation < 5% (~0.5dB) amplitude mismatch < 1mm distance mismatch

Sensitivity of Phase Offset Cancellation

81

• Rough prototype good for 802.15.4

• More precision needed for higher power systems (802.11)

Amplitude Mismatch between TX1 and TX2

Placement Errorfor RX

dB

Can

cella

tion

(dB)

Can

cella

tion

(dB)

Error (mm)

Sensitivity of Phase Offset Cancellation

Bandwidth Constraint

82

fc

d d + λ/2

TX1 TX2RX

A λ/2 offset is precise for one frequency

Bandwidth Constraint

A λ/2 offset is precise for one frequencynot for the whole bandwidth

83

fc fc+Bfc -B

d d + λ/2

TX1 TX2RX

Bandwidth Constraint

A λ/2 offset is precise for one frequencynot for the whole bandwidth

84

fc fc+Bfc -B

d d + λ/2

TX1 TX2RX

d2 d2 + λ+B/2

TX1 TX2RX

d1 d1 + λ-B/2

TX1 TX2RX

Bandwidth Constraint

A λ/2 offset is precise for one frequencynot for the whole bandwidth

85

fc fc+Bfc -B

d d + λ/2

TX1 TX2RX

d2 d2 + λ+B/2

TX1 TX2RX

d1 d1 + λ-B/2

TX1 TX2RX

WiFi (2.4G, 20MHz) => ~0.26mm precision error

Bandwidth Constraint

86

2.4 GHz

5.1 GHz

300 MHz

fc

Edge frequency

Bandwidth Constraint

87

• WiFi (2.4GHz, 20MHz): Max 47dB reduction

• Bandwidth⬆ => Cancellation⬇• Carrier Frequency⬆ => Cancellation⬆

2.4 GHz

5.1 GHz

300 MHz

Mitigating Hidden Terminals

88

0.5

0.6

0.7

0.8

0.9

0 2000 4000 6000 8000Data Load (Kbps)

Pack

et R

ecep

tion

Ratio

Full Duplex

Half Duplex

0.5

0.6

0.7

0.8

0.9

0 2000 4000 6000 8000

Mitigating Hidden Terminals

89

Data Load (Kbps)

Pack

et R

ecep

tion

Ratio

Full Duplex

Half Duplex

• Full-duplex reduces hidden terminal related losses by 88% at 2 Mbps

Mitigating Hidden Terminals

90

0.700

0.775

0.850

0.925

1.000

0 2000 4000 6000 8000Data Load (Kbps)

Fairn

ess

(JFI

)

Full Duplex

Half Duplex

• Full-duplex reduces hidden terminal related losses by 88% at 2 Mbps

• At higher loads, half-duplex improves PRR at the expense of fairness

0.5

0.6

0.7

0.8

0.9

0 2000 4000 6000 8000Data Load (Kbps)

Pack

et R

ecep

tion

Ratio

Full Duplex

Half Duplex

Time

91

Passive components better than active components

• No gain required

• Saturation can lead to non-linearity

• Passive components are more frequency flat

TX RX

TXRF Frontend

Attenuator andDelay Line

Xt

+Xt/2 -Xt/2

∑

RXRF Frontend

92

• ~65% converge without going through a local minima

• 98% converge in <20 iterations

93

AnalogConversionand Shaping

TX Signal TXFiltering and

DigitalConversion

RX

∑+-Channel

Model

DigitalReceiver

CancellationSignal

Digital Cancellation

Residual Self-interference

• Other cancellation techniques

Digital estimation for RF cancellation[1]

• Non-linear channel response

Reduce distortion: feedforward amplifiers

94

TX Signal

EstimateDistortion

∑+ -

High PowerAmplifier