scispace - formally typeset
Search or ask a question
Proceedings ArticleDOI

Practical, real-time, full duplex wireless

TL;DR: Experimental results show that a re- design of the wireless network stack to exploit full duplex capability can result in significant improvements in network performance.
Abstract: This paper presents a full duplex radio design using signal inversion and adaptive cancellation. Signal inversion uses a simple design based on a balanced/unbalanced (Balun) transformer. This new design, unlike prior work, supports wideband and high power systems. In theory, this new design has no limitation on bandwidth or power. In practice, we find that the signal inversion technique alone can cancel at least 45dB across a 40MHz bandwidth. Further, combining signal inversion cancellation with cancellation in the digital domain can reduce self-interference by up to 73dB for a 10MHz OFDM signal. This paper also presents a full duplex medium access control (MAC) design and evaluates it using a testbed of 5 prototype full duplex nodes. Full duplex reduces packet losses due to hidden terminals by up to 88%. Full duplex also mitigates unfair channel allocation in AP-based networks, increasing fairness from 0.85 to 0.98 while improving downlink throughput by 110% and uplink throughput by 15%. These experimental results show that a re- design of the wireless network stack to exploit full duplex capability can result in significant improvements in network performance.

Summary (4 min read)

1. INTRODUCTION

  • Wireless radios today are generally half duplex.
  • Antenna cancellation, when combined with the other mechanisms, allows full duplex operation: Choi et al. evaluate working 802.15.4 (1mW) full duplex prototypes which are close to (within 8%) an ideal full duplex system.
  • Antenna cancellation-based designs have three major limitations.
  • Furthermore, having two transmit antennas creates slight null regions of destructive interference in the far field.
  • While manual tuning is sufficient for lab experiments demonstrating a proof of concept, it leaves open the question of whether it is possible to build a full duplex system that can automatically adapt to realistic environments.

2. RADIO DESIGN AND FULL DUPLEX

  • This section provides background on the basics of radio design as well as full duplex.
  • It explains the basic challenges in building a full duplex radio, the existing techniques for doing so, and the limitations of those techniques.

2.1 Radio Design

  • Figure 1 shows the basic design of a modern radio receiver.
  • Because digitally sampling a 2.4GHz signal would require very high speed sampling at the Nyquist frequency of 4.8GHz, radios downconvert a RF sig- 1These channel sounders are wideband (∼240MHz) radios used for RF profiling, programmed to generate a single wideband pilot pattern for measuring the channel.
  • Nal to a baseband signal centered around 0Hz.
  • To transmit a packet, a radio generates digital samples for the desired waveform, converts them to a baseband signal with a digital to analog converter (DAC) and upconverts the baseband to RF.

2.2 Cancellation

  • The problem is that a node hears not only the signal it wants to receive, but also the signal it is transmitting.
  • By canceling this self-interference from what it receives, a full duplex node can in theory decode a received signal.
  • Digital cancellation, while helpful, is by itself insufficient: current systems in the research literature cancel up to 20-25 dB [8, 7].
  • The limitation is that ADCs have a limited dynamic range: since self-interference is extremely strong, an ADC can quantize away the received signal, making it unrecoverable after digital sampling.
  • One approach to analog cancellation uses a second transmit chain to create an analog cancellation signal from a digital estimate of the self-interference [5], canceling ≈33dB of self-interference over a 625kHz bandwidth signal.

2.3 Full Duplex

  • Motivated by the above limitations, recent work has proposed antenna placement techniques [4, 5].
  • The separation between the receive and transmit antennas attenuates the self-interference signal [5], but this antenna separation is not enough.
  • The key idea behind antenna cancellation is to use a second transmit antenna and place it such that the two transmit signals interfere destructively at the receive antenna.
  • Combined with the analog and digital cancellation described above, the design is able to cancel 50-60dB of self-interference, which is sufficient to operate a full duplex 802.15.4 radio.
  • The third limitation relates to the details of Choi et al.’s design.

3.1 Eliminating the Bandwidth Constraint

  • The design is based on a simple observation: any radio that inverts a signal through adjusting phase will always encounter a bandwidth constraint that bounds its maximum cancellation.
  • The key insight of this paper is that one can use a balun in a completely new way, to obtain the inverse of a self-interference signal and use the inverted signal to cancel the interference.
  • To cancel selfinterference, the radio combines the negative signal with its received signal after adjusting the delay and attenuation of the negative signal to match the self-interference.
  • Nal before combining it, which may be hard to achieve in practice.
  • The rest of this section examines these practical concerns and how to solve them.

3.2 Balun Benefits

  • The first wire is an ideal selfinterference path and has a 20dB attenuator representing the antenna separation.
  • An RF combiner adds the two signals on the received side to measure the canceled signal.
  • In both cases, the passive delay line and attenuator provide fine-grained control to match phase and amplitude for the interference and cancellation paths to maximize cancellation.
  • In comparison, the balun based circuit provides a good degree of cancellation over a much wider bandwidth.
  • Based on Figure 5, the authors can obtain the best possible cancellation with balun and phase-offset cancellation for a given signal bandwidth.

3.3 Auto-tuning

  • The results in Figure 6 show that, if the phase and amplitude of the inverted signal are set correctly, balun cancellation can have impressive results across a wide bandwidth.
  • Since the QHx220 is an active component, the authors call this cancellation scheme Balun Active Cancellation.
  • It emulates a variable delay by controlling the attenuation of the in-phase and quadrature signals (gi and gq), adding them to create the output.
  • Hence the authors can use the same gradient descent algorithm for tuning the two attenuation factors in QHx220, gi and gq .
  • It then reverses direction, reduces the step size and attempts to converge to the optimal point.

3.4 Digital Cancellation

  • But this cancellation only handles the dominant self-interference component between the receive and transmit antennas.
  • The full duplex radio design uses digital cancellation (DC) to cancel any residual interference that persists after balun cancellation.
  • The estimation uses the least square algorithm [15] due to its low complexity.
  • Let s[n] be the known transmitted digital sample at time n fed into the FIR filter.
  • A node needs to re-estimate its channel state at a period below the channel coherence time.

4. FULL DUPLEX MAC

  • The most interesting possible benefits of full duplex occur above the physical layer.
  • Prior work, for example, argued that MAC layer techniques that exploit full duplex can mitigate hidden terminals and improve fairness [4, 13].
  • The lack of a real-time full duplex MAC layer prevented experimentally evaluating these claims.
  • The authors describe the design and implementation of a simple full duplex MAC.
  • Later sections evaluate a real-time full duplex network using this MAC.

4.1 Design

  • In CSMA/CA, the hidden terminal problem occurs because a half-duplex receiver cannot inform other nodes of an ongoing reception.
  • Full duplexing naturally mitigates the hidden terminal problem because the receiver can immediately start a transmission that also suppresses nearby nodes.
  • Furthermore, when a WLAN network is saturated, a bottleneck occurs at the AP because it gets the same transmit opportunity as each client while serving more flows.
  • The authors MAC uses busytones as a way to mitigate this problem.
  • Whenever a node finishes transmitting a packet before finishing receiving, it transmits a predefined signal until its reception ends.

4.2 Real-Time

  • The previous section has described the design of a full duplex MAC protocol.
  • The earlier the destination address is in the packet and the faster the hardware can initiate a secondary transmission, the lower the secondary response latency becomes.
  • Typical software-defined radios such as USRP [2] cannot meet these requirements, due to latency between hardware and the host PC as well as their need to store packets as digital samples rather than bits.
  • Each OFDM symbol is 8µs long, thus QPSK modulation achieves 12Mbps bitrate without any channel coding.
  • Primary transmissions use the existing half duplex MAC implementation from the same reference design which mimics 802.11 behavior.

5. EVALUATION

  • This section provides experimental results for the performance of the adaptive full duplex system described in Sections 3 and 4.
  • The evaluation has two parts: physical layer performance of the cancellation design and effect of a full duplex MAC on hidden terminals and network fairness.

5.1 Cancellation

  • The authors evaluate the cancellation design by measuring how much it can attenuate the self-interference signal.
  • As mentioned in Section 3.3.1, active components like QHx220 introduce non-linearities which limit the efficacy of balun and digital cancellation.
  • Figure 11 shows the cancellation achieved with the combination of balun active and digital cancellation as the transmit power of the self-interfering signal varies.
  • Each data point takes 26µs, 10µs to change the QHx220 settings and 16µs to measure RSSI over two OFDM symbols.
  • Figure 13 shows the cumulative distribution of how many iterations the algorithm takes to converge.

5.2 MAC performance

  • The previous section proves full duplex feasibility with real-time self interference cancellation over wideband WiFi signals.
  • Specifically the authors focus on the hidden terminal and fairness problems.
  • The reason is that the data load of each flow is 3Mbps, which is lower than the link capacity of around 8Mbps.
  • Full duplex should increase the overall throughput by a factor of two, while the results show only a 45% overall increase.
  • Due to bursty traffic, sometimes the queue at the AP does not have packets for all clients.

6. COMPARISON WITH MIMO

  • The performance evaluation in the previous section shows the clear benefits of the full duplex in the MAC performance.
  • The authors compare the case when the nodes use the two antennas to implement a 2x2 MIMO system vs implement a full duplex system.
  • Without channel state, the transmitter cannot change its rate in response to channel changes and should fix its rate in advance.
  • This section shows that for different channel conditions the performance of full duplex can exceed or lag behind the performance of a similarly resourced 2x2 MIMO system.
  • If so, this raises an interesting and new degree of freedom.

7. DISCUSSION

  • A balun cancellation-based full duplex radio, while an important step forward, leaves several open questions and new possibilities and future research.
  • One issue that comes up numerous times in the design is the sensitivity and precision of the components used in cancellation.
  • These accuracies are clearly possible – a small group of researchers were able to achieve them with commodity components.
  • Using in-packet techniques to update channel estimates on a per-packet basis can address this challenge.
  • While the full duplex system presented does not preclude coexistence with existing half duplex systems, secondary transmissions need to know whether the primary is full duplex capable, otherwise there may be poor interactions with link layer retry counts.

8. CONCLUSION

  • This paper presents design for a full duplex radio based on a novel cancellation technique called balun cancellation.
  • Unlike prior designs, balun cancellation can work on wideband, high power signals, such that it is possible to build full duplex 802.11n devices.
  • These devices can automatically tune their cancellation circuits and so operate in dynamic environments.
  • The paper describes the implementation of a full duplex prototype that uses an 802.11-like OFDM link layer which can operate in real time.
  • Using this prototype, it evaluates a full duplex MAC layer, validating prior hypotheses that full duplex can prevent many hidden terminals and improve wireless LAN fairness.

Did you find this useful? Give us your feedback

Figures (16)

Content maybe subject to copyright    Report

Practical, Real-time, Full Duplex Wireless
Mayank Jain
1
, Jung Il Choi
1
, Tae Min Kim
1
, Dinesh Bharadia
1
, Siddharth Seth
1
,
Kannan Srinivasan
2
, Philip Levis
1
, Sachin Katti
1
, Prasun Sinha
3
1
Stanford University
2
The University of Texas at Austin
3
The Ohio State University
California, USA Texas, USA Ohio, USA
{mayjain, jungilchoi, kimtm, dineshb, sidseth}@stanford.edu,
kannan.srinivasan@austin.utexas.edu, pal@cs.stanford.edu,
skatti@stanford.edu, prasun@cse.ohio-state.edu
Co-primary authors
Abstract
This paper presents a full duplex radio design using signal inversion
and adaptive cancellation. Signal inversion uses a simple design
based on a balanced/unbalanced (Balun) transformer. This new de-
sign, unlike prior work, supports wideband and high power sys-
tems. In theory, this new design has no limitation on bandwidth or
power. In practice, we find that the signal inversion technique alone
can cancel at least 45dB across a 40MHz bandwidth. Further, com-
bining signal inversion cancellation with cancellation in the digital
domain can reduce self-interference by up to 73dB for a 10MHz
OFDM signal.
This paper also presents a full duplex medium access control
(MAC) design and evaluates it using a testbed of 5 prototype full
duplex nodes. Full duplex reduces packet losses due to hidden ter-
minals by up to 88%. Full duplex also mitigates unfair channel
allocation in AP-based networks, increasing fairness from 0.85 to
0.98 while improving downlink throughput by 110% and uplink
throughput by 15%. These experimental results show that a re-
design of the wireless network stack to exploit full duplex capabil-
ity can result in significant improvements in network performance.
Categories and Subject Descriptors
C.2.1 [Computer-Communication Networks]: Network Archi-
tecture and Design—Wireless communication
General Terms
Design, Performance, Reliability
Keywords
Full-duplex wireless
1. INTRODUCTION
Wireless radios today are generally half duplex. On a single
channel, they can either transmit or receive, but not both simultane-
Permission to make digital or hard copies of all or part of this work for
personal or classroom use is granted without fee provided that copies are
not made or distributed for profit or commercial advantage and that copies
bear this notice and the full citation on the first page. To copy otherwise, to
republish, to post on servers or to redistribute to lists, requires prior specific
permission and/or a fee.
MobiCom’11, September 19–23, 2011, Las Vegas, Nevada, USA.
Copyright 2011 ACM 978-1-4503-0492-4/11/09 ...$10.00.
ously. The inability to do both simultaneously is the root of many
of the open problems in wireless today, such as media access con-
trol and bitrate adaptation. A full duplex radio has the potential to
completely reconsider how we design and build wireless networks.
The possible benefits of full duplex wireless have recently led
researchers to explore how one might build such a device [4]. The
basic challenge is reducing self interference. If a node cannot hear
its own signal, then its own transmissions will not interfere with
other packets: it can simultaneously transmit and receive.
Well known digital and analog techniques, even when combined
together, are not sufficient to cancel self-interference sufficiently
for full duplex [4]. Some analog cancellation techniques use noise
canceling chips to subtract the self transmission signal (the “noise”)
from the received signal [12]. Digital cancellation, used in CS-
MA/CN, optical networks and proposals for full duplex operation,
subtracts self-interference in the digital domain, after the receiver
has converted the baseband signal to digital samples [14, 6, 3].
Motivated by these limitations, recent work has explored antenna
placement as an additional cancellation technique. Antenna separa-
tion uses the fact that the distance between the transmit and receive
antennas naturally reduces self-interference due to signal attenu-
ation [5]. However, we would need impractically large distances
between the TX and RX antennas to obtain enough reduction to
make full duplex possible with only antenna separation. To further
cancel self-interference, Choi et al. propose an additional tech-
nique, called antenna cancellation [4]. Antenna cancellation, when
combined with the other mechanisms, allows full duplex operation:
Choi et al. evaluate working 802.15.4 (1mW) full duplex proto-
types which are close to (within 8%) an ideal full duplex system.
Although promising, antenna cancellation-based designs have
three major limitations. The first is that they require three antennas:
two transmit, and one receive. Full duplex doubles throughput, but
with three antennas a MIMO system can triple throughput, so from
a raw performance standpoint antenna cancellation is unattractive.
Furthermore, having two transmit antennas creates slight null re-
gions of destructive interference in the far field. The second lim-
itation is a “bandwidth constraint, a theoretical limit which pre-
vents supporting wideband signals such as WiFi. Finally, Choi et
al.s design introduces a third, practical limitation: it requires man-
ual tuning. While manual tuning is sufficient for lab experiments
demonstrating a proof of concept, it leaves open the question of
whether it is possible to build a full duplex system that can auto-
matically adapt to realistic environments.

Baseband
Mapping
0011100
Transmit
Bits
RF Mixer
Baseband
Signal
RF TX
Signal
TX
Carrier
Frequency
Baseband
Samples
Digital to
Analog
Converter
(DAC)
Baseband
Demapping
0011100
Receive
Bits
RF Mixer
Baseband
Signal
RF RX
Signal
RX
Carrier
Frequency
Baseband
Samples
Analog to
Digital
Converter
(ADC)
Figure 1: Simplified block diagram of an RF Receiver
The first contribution of this paper is a full duplex radio design
that addresses all three limitations. It proposes a novel mechanism,
balun cancellation, that uses signal inversion, through a balun (bal-
anced/unbalanced) circuit. Balun cancellation has no bandwidth
constraint: it can in theory support arbitrary bandwidths and can-
cel arbitrarily high transmit powers. Balun cancellation requires
only two antennas, one transmit and one receive. Furthermore, this
paper presents a tuning algorithm that allows a balun-based radio
design to quickly, accurately, and automatically adapt the full du-
plex circuitry to cancel the primary self-interference component.
The second contribution is the implementation and evaluation of
the balun cancellation design as well as a full duplex MAC proto-
col. It explores how and where practice differs from theory, exam-
ining the engineering challenges that arise in making full duplex
practical. For example, while a balun can in theory provide perfect
cancellation, this assumes that the balun has a flat frequency re-
sponse: if not flat, the inverted signal might differ slightly from the
transmitted signal. These results indicate possible future challenges
in large scale full duplex radio production.
We evaluate the design in two ways. The first is using a tightly
controlled channel sounder
1
to demonstrate the limits of the exist-
ing circuitry at the physical layer. We find that well-tuned balun
cancellation circuit built from commodity components can cancel
over 45dB across a 40MHz bandwidth. Combined with digital
cancellation, this allows a full duplex radio to cancel up to 73dB
of self-interference: consequently full duplex 802.11n devices are
possible with a reasonable separation between TX and RX anten-
nas. The second evaluation uses a 5-node WARP software radio
testbed to quantify the link-layer benefits of a full duplex MAC
layer based on the Contraflow design [13].
2. RADIO DESIGN AND FULL DUPLEX
This section provides background on the basics of radio design
as well as full duplex. It explains the basic challenges in building
a full duplex radio, the existing techniques for doing so, and the
limitations of those techniques.
2.1 Radio Design
Figure 1 shows the basic design of a modern radio receiver. We
walk through these details because the underlying data representa-
tions determine how and when a full duplex radio can cancel sig-
nals. We use channel 1 of 802.11b as a running example to ground
the concepts in concrete numbers.
A wireless signal occupies a bandwidth, a range of frequen-
cies. 802.11b channels, for example, are 22MHz wide. Chan-
nel 1 of 802.11b is centered at 2.412GHz: it spans 2.401GHz to
2.423GHz. The signal transmitted and received at this frequency
range is called the RF (Radio Frequency) signal. Because digitally
sampling a 2.4GHz signal would require very high speed sampling
at the Nyquist frequency of 4.8GHz, radios downconvert a RF sig-
1
These channel sounders are wideband (240MHz) radios used
for RF profiling, programmed to generate a single wideband pilot
pattern for measuring the channel.
Digital
Cancellation
d
d + λ/2
TX1 TX2RX
ADC
RF Interference
Cancellation
Antenna Cancellation
Tx signal
Figure 2: Block diagram of existing full duplex design with
three cancellation techniques.
nal to a baseband signal centered around 0Hz. The baseband signal
of 802.11b occupies -11 to 11 MHz. Downconverting allows the ra-
dio to use a much lower speed analog to digital converter (ADC):
the 22MHz baseband signal needs an ADC operating at or slightly
above the Nyquist rate of 44 MHz. Commodity WiFi cards typi-
cally use 8-bit samples, though some software radios can provide
12 bit resolution.
To transmit a packet, a radio generates digital samples for the
desired waveform, converts them to a baseband signal with a digital
to analog converter (DAC) and upconverts the baseband to RF. To
receive a packet, a radio downconverts an RF signal to a baseband
signal, then samples the baseband with an ADC to generate digital
samples.
2.2 Cancellation
The goal of a full duplex radio is to transmit and receive simul-
taneously. The problem is that a node hears not only the signal it
wants to receive, but also the signal it is transmitting. By cancel-
ing this self-interference from what it receives, a full duplex node
can in theory decode a received signal. This cancellation is hard,
however, because the self-interference can be millions to billions
of times stronger (60-90 dB) than a received signal. For example, a
radio with a transmit power of 0dBm and a noise floor of approxi-
mately -90dBm needs to cancel nearly 95dB of self-interference to
ensure that its own transmissions do not disrupt reception.
There are existing digital and analog techniques to cancel inter-
ference. Digital cancellation operates on digital samples. If a full
duplex radio has a good estimate of the phase and amplitude of its
transmitted signal at the receive antenna, it can generate the dig-
ital samples for its transmitted signal and subtract them from its
received samples. Digital cancellation, while helpful, is by itself
insufficient: current systems in the research literature cancel up to
20-25 dB [8, 7]. The limitation is that ADCs have a limited dy-
namic range: since self-interference is extremely strong, an ADC
can quantize away the received signal, making it unrecoverable af-
ter digital sampling.
Analog cancellation uses knowledge of the transmission to can-
cel self-interference in the RF signal, before it is digitized. One
approach to analog cancellation uses a second transmit chain to
create an analog cancellation signal from a digital estimate of the
self-interference [5], canceling 33dB of self-interference over a
625kHz bandwidth signal. Another approach uses techniques sim-
ilar to noise-canceling headphones [12]. The self-interference sig-
nal is the “noise” which a circuit the QHx220 chip [11] sub-
tracts from the received signal. The QHx220 can cancel 20-25 dB
for a 10MHz bandwidth signal but it introduces numerous com-
plications, such as non-linearities and distortions, which compli-
cate digital cancellation significantly. We examine these limitations
more deeply in Section 5.1, but the overall result is that this ap-

proach cannot provide more than 25 dB of cancellation and cannot
be combined with digital cancellation: it is therefore insufficient
for full duplex.
2.3 Full Duplex
Motivated by the above limitations, recent work has proposed
antenna placement techniques [4, 5]. The state of the art in full du-
plex operates on narrowband 5MHz signals with a transmit power
of 0dBm (1mW) [4]. The design achieves this result by augmenting
the digital and analog cancellation schemes described above with a
novel form of cancellation called “antenna” cancellation. Figure 2
shows a block diagram of this design. The separation between
the receive and transmit antennas attenuates the self-interference
signal [5], but this antenna separation is not enough. The key
idea behind antenna cancellation is to use a second transmit an-
tenna and place it such that the two transmit signals interfere de-
structively at the receive antenna. It achieves this by having one-
half wavelength distance offset between the two transmit anten-
nas. The resulting destructive interference cancels 20-30dB of self-
interference. Combined with the analog and digital cancellation de-
scribed above, the design is able to cancel 50-60dB of self-interference,
which is sufficient to operate a full duplex 802.15.4 radio.
This design, while a breakthrough, has both fundamental and
practical limitations. The first limitation is fundamental and re-
lates to the bandwidth of the transmitted signal. Antenna cancel-
lation ensures that only the signal at the center frequency is per-
fectly inverted in phase at the receive antenna and thus cancelled
fully. However, as a signal moves further away from the center
frequency, the phase offset of two versions from the transmit an-
tennas shifts away from perfect inversion and so the two do not
cancel completely. Correspondingly, antenna cancellation perfor-
mance degrades as the bandwidth of the signal to cancel increases.
In practice, this means antenna cancellation cannot cancel an
802.11b signal by more than 30dB: Choi et al. note that it may be
just barely possible to build a full duplex 802.11b node [4]. Widen-
ing the signal to 40MHz (e.g., 802.11n) precludes full duplex at
WiFi transmit powers.
Furthermore, as the cancellation is highly frequency selective,
modulation approaches such as OFDM which break a bandwidth
into many smaller parallel channels will perform even more poorly.
Due to frequency selectivity, different subcarriers will experience
drastically different self-interference. Hence cancellation suffers
for wideband OFDM systems that do not adapt on a per subcar-
rier basis. As WiFi and other wireless technologies move towards
OFDM (802.11n and 802.11g use it) due to its technical advan-
tages, antenna cancellation will encounter serious challenges.
Antenna cancellation’s second limitation is requiring three an-
tennas. Full duplex at most doubles throughput. But with three
antennas, a 3x3 MIMO array can theoretically triple throughput.
This comparison suggests it may be better to just use MIMO and
sacrifice some portion of the available capacity for other useful link
layer primitives such as RTS/CTS. Furthermore, having two trans-
mit antennas creates slight null regions of destructive interference
in the far field.
The third limitation relates to the details of Choi et al.s design.
The full duplex radio requires manually tuning the phase and am-
plitude of the second transmit antenna to maximize cancellation at
the receive antenna. While manual tuning is sufficient for a lab
proof-of-concept, it is an open question whether it is technically
feasible to design a circuit that automatically tunes itself in a dy-
namic environment.
Digital Interference Cancellation
TX
RX
Attenuation
& Delay
RF ! Baseband
ADC
Baseband ! RF
DAC
Encoder
Decoder
Digital Interference
Reference
Balun Cancellation
TX Signal Path RX Signal Path
RF Reference
Σ
FIR filter
-
RSSI
Control
Feedback
Channel
Estimate
Balun
Figure 3: Block diagram of full duplex system. The ideal can-
cellation setup uses passive, high precision components for at-
tenuation and delay adjustment.
3. DESIGN
In this section we describe the design of a full duplex system
radio that requires only 2 antennas and RF frontends, has no band-
width constraint, and automatically tunes its self-interference can-
cellation in response to changing channel conditions.
3.1 Eliminating the Bandwidth Constraint
The design is based on a simple observation: any radio that
inverts a signal through adjusting phase will always encounter a
bandwidth constraint that bounds its maximum cancellation. To
cancel beyond this bound, a radio needs to obtain the perfect in-
verse of a signal, i.e., a signal which is the perfect negative of the
transmitted signal at all instants. Combining this inverse with the
transmitted signal can in theory completely cancel self-interference.
All a radio needs is to invert a signal without adjusting its phase.
Luckily, there is a component that does exactly that: a balanced/un-
balanced (balun) transformer. Baluns are a common component in
RF, audio and video circuits for converting back and forth between
single-ended signals single-wire signals with a common ground
and differential signals two-wire signals with opposite polarity.
For example, converting a single-ended signal on a co-axial cable
to a differential signal for transmission on a twisted pair cable (such
as Ethernet), or vice-versa, uses a balun to take the signal as input
and output the signal and its inverse.
The key insight of this paper is that one can use a balun in a
completely new way, to obtain the inverse of a self-interference
signal and use the inverted signal to cancel the interference. Fig-
ure 3 shows a 2-antenna full duplex radio design that uses a balun.
The transmit antenna transmits the positive signal. To cancel self-
interference, the radio combines the negative signal with its re-
ceived signal after adjusting the delay and attenuation of the nega-
tive signal to match the self-interference.
We call this technique balun passive cancellation since we ide-
ally use high precision passive components to realize the variable
attenuation and delay in the cancellation path. While balun cancel-
lation can in theory cancel perfectly, there are of course practical
limitations. For example, the transmitted signal on the air expe-
riences attenuation and delay. To obtain perfect cancellation the
radio must apply identical attenuation and delay to the inverted sig-

240 MHz TX
@2.4GHz
RX
Variable
Attenuator
Balun
Variable
Delay Line
!
20dB
Attenuator
Self-interference path
Cancellation Path
Figure 4: Wired setup to measure the cancellation performance
of signal inversion vs phase offset. The phase offset experiment
uses an RF splitter instead of a balun to split the signal.
2300 2350 2400 2450 2500 2550 2600
−110
−100
−90
−80
−70
−60
Frequency in Mhz
Received signal (dBm)
2300 2350 2400 2450 2500 2550 2600
−110
−100
−90
−80
−70
−60
Balun
Phase Offset
Figure 5: Cancellation of the self-interference signal with the
balun vs with phase offset. The received signal is -49dBm with-
out any cancellation. Using a balun gives a flatter cancellation
response.
nal before combining it, which may be hard to achieve in practice.
Moreover, the balun may have engineering imperfections, such as
leakage or a non-flat frequency response. The rest of this section
examines these practical concerns and how to solve them.
3.2 Balun Benefits
To understand the practical benefits and limitations of inverting
a signal with a balun, compared to inverting it with a phase offset,
we conduct a tightly controlled RF experiment. Figure 4 shows
the experimental setup. We program a signal generator to generate
a wideband 240MHz chirp with a center frequency of 2.45GHz.
This signal goes over two wires. The first wire is an ideal self-
interference path and has a 20dB attenuator representing the an-
tenna separation. The second wire goes through a cancellation path,
consisting of a variable attenuator and variable delay element that
can be controlled to modify the cancellation path signal to match
the self-interference. The combination of the two signals feeds into
a signal receiver. The variable delay line and attenuator in this ex-
periment are manually tunable passive devices which allow for a
high degree of precision.
We simulate antenna cancellation by making the cancellation
path one half of a wavelength longer than the self-interference path.
An RF combiner adds the two signals on the received side to mea-
sure the canceled signal. The balun setup, on the other hand, uses a
balun to split the transmit signal, and uses wires of the same length
for the self-interference and cancellation paths. In both cases, the
passive delay line and attenuator provide fine-grained control to
match phase and amplitude for the interference and cancellation
paths to maximize cancellation.
Figure 5 shows the results. Using a phase-offset signal can-
cels well over a narrow bandwidth, but is very limited in canceling
wideband signals. Phase offset cancellation can cancel 50dB for a
5MHz signal, but only provide 25dB of cancellation for a 100MHz
0 20 40 60 80 100 120
20
30
40
50
60
Bandwidth in Mhz
Cancellation (dB)
0 20 40 60 80 100 120
20
30
40
50
60
Phase Offset
Balun
Figure 6: Cancellation performance with increasing signal
bandwidth when using the balun method vs using phase offset
cancellation.
signal. In comparison, the balun based circuit provides a good de-
gree of cancellation over a much wider bandwidth. For example,
balun based cancellation would provide 52dB of cancellation for a
5MHz signal and 40dB of cancellation for a 100MHz signal.
Balun cancellation is not perfect across the entire band. The key
reason is that the balun circuit is not frequency flat, i.e., different
parts of the band are inverted with different amplitudes. Conse-
quently applying a single attenuation and delay factor to the in-
verted signal will not cancel the transmitted signal perfectly: this
is a simple instance of real-world engineering tolerances limiting
theory. Based on Figure 5, we can obtain the best possible cancel-
lation with balun and phase-offset cancellation for a given signal
bandwidth. Figure 6 shows the best cancellation achieved using
each method for signals of varying bandwidths.
3.3 Auto-tuning
The results in Figure 6 show that, if the phase and amplitude of
the inverted signal are set correctly, balun cancellation can have
impressive results across a wide bandwidth. This raises a sim-
ple follow-on question. Is it possible to automatically adjust the
phase and amplitude, thereby self-tuning cancellation in response
to channel changes? In this section we describe an algorithm that
can accurately and quickly self-tune a cancellation circuit.
The basic approach is to estimate the attenuation and delay of the
self-interference signal and match the inverse signal appropriately.
Ideally, the auto-tuning algorithm would adjust the attenuation and
delay such that the residual energy after balun cancellation would
be minimized (assuming no other signal is being received on the
RX antenna). Let g and τ be the variable attenuation and delay
factors respectively, and s(t) be the signal received at the input of
the programmable delay and attenuation circuit. The delay over
the air (wireless channel) relative to the programmable delay is τ
a
.
The attenuation over the wireless channel is g
a
. The energy of the
residual signal after balun cancellation is:
E =
Z
T
o
(g
a
s(t τ
a
) gs(t τ))
2
dt (1)
where T
o
is the baseband symbol duration. The goal of the algo-
rithm is to adjust the parameters g and τ to minimize the energy of
the residual signal.
Our insight is that the residual energy function in Eq. 1 has a
pseudo-convex relationship with g and τ for WiFi style OFDM sig-
nals. We omit the mathematics for brevity but note that we can ex-
ploit this structure to design a simple gradient descent algorithm to
converge to the optimal delay and attenuation.
3.3.1 Practical Algorithm with QHx220
While a gradient descent algorithm would work well, fine-grained

TX
RX
gain
g
i
Output
Signal to
Transmit
RF Interference Cancellation
Circuit
+
-
Balun
Air Signal
+
RSSI
Wire Signal
Algorithm
Fixed
Delay(τ)
Control g
i,
g
q
+
gain
g
q
Qhx220
Figure 7: Block diagram of full duplex system with Balun Ac-
tive Cancellation. The RSSI values represent the energy re-
maining after cancellation. Our algorithm adapts attenuation
parameters (g
i
and g
q
) to minimize this energy.
programmable analog attenuation and delay lines are unfortunately
not typical commodity components and so are expensive. Hence
in our current implementation we have to make do with a compo-
nent that provides an approximation, the QHx220 noise cancella-
tion chip [11], which prior full duplex designs have used [4, 12].
Since the QHx220 is an active component, we call this cancellation
scheme Balun Active Cancellation.
Figure 7 shows the block diagram of Balun Active Cancellation
with the auto-tuning circuit. The RSSI value provides the residual
signal energy after balun cancellation has subtracted self interfer-
ence from the received signal. As we can see, the QHx220 does
not actually provide a variable delay. Instead, it takes the input
signal and separates it into an in-phase and quadrature component.
The quadrature component has a fixed delay (τ ) with respect to the
in-phase component. It emulates a variable delay by controlling
the attenuation of the in-phase and quadrature signals (g
i
and g
q
),
adding them to create the output.
For a single frequency, this approach can correctly emulate any
phase. However, for signals with a bandwidth, the fixed delay τ
only matches one frequency. This fixed delay suffers from a similar
bandwidth constraint as antenna cancellation. Prior work also dis-
cusses the limitation of the cancellation model used by QHx220 [9].
The goal of the auto-tuning algorithm in this case would be to
find the attenuation factors on both lines such that the QHx220 chip
output is the best approximation of the self interference we need
to cancel from the received signal. Fortunately, we can show that
even with this approximate version, we still retain a psuedo-convex
structure. To see this empirically, we conduct an experiment where
the TX antenna transmits a 10MHz OFDM signal, and we vary the
two attenuations in QHx220. We plot the RSSI output in Figure 8,
where a deep null exists at the optimal point. Hence we can use
the same gradient descent algorithm for tuning the two attenuation
factors in QHx220, g
i
and g
q
. The algorithm works in steps, and
at each step it computes the slope of the residual RSSI curve by
changing g
i
and g
q
by a fixed step size. If the new residual RSSI is
lower than before, then it moves to the new settings for the atten-
uation factors, and repeats the process. If at any point it finds that
the residual RSSI increases, it knows that it is close to the optimal
point. It then reverses direction, reduces the step size and attempts
to converge to the optimal point. The algorithm also checks for
false positives, caused due to noisy minimas. The algorithm is fast;
it typically converges to the minimum in 8 15 iterations, depend-
ing on the choice of the starting point. An iteration requires 5 mea-
surements, each involving a phase/attenuation adjustment followed
by RSSI sampling.
Figure 8: RSSI of the residual signal after balun cancellation
as we vary g
i
and g
q
in the QHx220. Note the deep null at the
optimal point.
Caveats: The commodity parts of our prototype introduce a few
limitations. The QHx220 has hardware constraints, such that it
causes non-linear distortion, especially for input powers beyond -
40dBm. Hence, cancellation will not be perfect for typical wireless
input powers (0-30dBm). Non-linear distortion also impacts digital
cancellation, as we see in Section 5.1.
Our current implementation uses the QHx220 despite its imper-
fections because it is inexpensive and easily available. However,
we believe that it is feasible to build a full duplex radio using an
electronically tunable delay and attenuation chipset, since they are
commercially available but just not widely and inexpensively. Fur-
thermore, including them as small parts of a full duplex radio hard-
ware design would not be particularly complex or expensive.
3.4 Digital Cancellation
As Figure 5 shows, a industry-grade balun cancellation circuit
can cancel up to 45dB of a 40MHz signal. But this cancellation
only handles the dominant self-interference component between
the receive and transmit antennas. A node’s self-interference may
have other multipath components, which, although much weaker
than the dominant one, are strong enough to interfere with recep-
tion. Furthermore, the balun circuit may distort the cancellation
signal slightly, such that it introduces some interference leakage.
The full duplex radio design uses digital cancellation (DC) to
cancel any residual interference that persists after balun cancella-
tion. Implementing DC for a full duplex radio, however, is more
challenging than other uses of digital cancellation, such as succes-
sive interference cancellation (SIC) [8] and ZigZag decoding [7].
Unlike SIC or ZigZag, which use DC to recover packets which
would have otherwise been lost, a full duplex radio uses DC to pre-
vent the loss of packets which a half duplex radio could receive.
While an SIC implementation that recovers 80% of otherwise lost
packets is a tremendous success, a full duplex radio that drops 20%
of packets is barely usable.
To the best of our knowledge, our DC system has three novel
achievements compared to existing software radio implementations
in the literature. First, it is the first real-time cancellation imple-
mentation that runs in hardware: this is necessary for the MAC
experiments in Section 5.2. Second, it is the first cancellation im-
plementation that can operate on 10MHz signals. Finally, it is the
first digital cancellation that operates on OFDM signals. For these
reasons, we give a very detailed description of its design and algo-
rithms.
Digital cancellation has two components: estimating the self-
interference channel; and using the channel estimate on the known
transmit signal to generate digital samples to subtract from the re-
ceived signal.

Citations
More filters
Proceedings ArticleDOI
27 Aug 2013
TL;DR: The design and implementation of the first in-band full duplex WiFi radios that can simultaneously transmit and receive on the same channel using standard WiFi 802.11ac PHYs are presented and achieves close to the theoretical doubling of throughput in all practical deployment scenarios.
Abstract: This paper presents the design and implementation of the first in-band full duplex WiFi radios that can simultaneously transmit and receive on the same channel using standard WiFi 802.11ac PHYs and achieves close to the theoretical doubling of throughput in all practical deployment scenarios. Our design uses a single antenna for simultaneous TX/RX (i.e., the same resources as a standard half duplex system). We also propose novel analog and digital cancellation techniques that cancel the self interference to the receiver noise floor, and therefore ensure that there is no degradation to the received signal. We prototype our design by building our own analog circuit boards and integrating them with a fully WiFi-PHY compatible software radio implementation. We show experimentally that our design works robustly in noisy indoor environments, and provides close to the expected theoretical doubling of throughput in practice.

2,084 citations


Cites background or methods from "Practical, real-time, full duplex w..."

  • ...Categories and Subject Descriptors C.2.1 [Computer Communication Networks]: Network Architecture and Design—Wireless communication General Terms: Algorithms, Design, Experimentation, Performance Keywords: Full Duplex, Interference Cancellation, Non-linear Cancellation...

    [...]

  • ...Designs which use an extra transmitter chain report an overall total of 80dB of self-interference cancellation (we have been able to reproduce their results experimentally)....

    [...]

  • ...…suggests that any in-band full duplex system has to be able to cancel all the above distortions in addition to the main signal component itself, since all of these are within the frequency band we are transmitting and receiving on and act as strong self-interference to the received signal itself....

    [...]

  • ...• A full duplex system has to reduce non-linear harmonic components that are 80dB above the noise floor, so any full duplex technique has to provide at least 80dB of non-linear self-interference cancellation....

    [...]

Journal ArticleDOI
TL;DR: In this article, the authors present a survey of self-interference mitigation techniques for in-band full-duplex (IBFD) wireless systems and discuss the challenges and opportunities in the design and analysis of IBFD wireless systems.
Abstract: In-band full-duplex (IBFD) operation has emerged as an attractive solution for increasing the throughput of wireless communication systems and networks. With IBFD, a wireless terminal is allowed to transmit and receive simultaneously in the same frequency band. This tutorial paper reviews the main concepts of IBFD wireless. One of the biggest practical impediments to IBFD operation is the presence of self-interference, i.e., the interference that the modem's transmitter causes to its own receiver. This tutorial surveys a wide range of IBFD self-interference mitigation techniques. Also discussed are numerous other research challenges and opportunities in the design and analysis of IBFD wireless systems.

1,752 citations

Posted Content
TL;DR: This tutorial surveys a wide range of IBFD self-interference mitigation techniques and discusses numerous other research challenges and opportunities in the design and analysis of IB FD wireless systems.
Abstract: In-band full-duplex (IBFD) operation has emerged as an attractive solution for increasing the throughput of wireless communication systems and networks. With IBFD, a wireless terminal is allowed to transmit and receive simultaneously in the same frequency band. This tutorial paper reviews the main concepts of IBFD wireless. Because one the biggest practical impediments to IBFD operation is the presence of self-interference, i.e., the interference caused by an IBFD node's own transmissions to its desired receptions, this tutorial surveys a wide range of IBFD self-interference mitigation techniques. Also discussed are numerous other research challenges and opportunities in the design and analysis of IBFD wireless systems.

1,549 citations


Cites methods from "Practical, real-time, full duplex w..."

  • ...5...

    [...]

  • ...In bidirectional IBFD, two modems simultaneously exchange messages over the same frequency band, as further described in Section III....

    [...]

  • ...14...

    [...]

  • ...10...

    [...]

Journal ArticleDOI
TL;DR: This work shows that the average amount of cancellation increases for active cancellation techniques as the received self-interference power increases, and shows that digital cancellation is more effective when applied selectively based on measured suppression values.
Abstract: We present an experiment-based characterization of passive suppression and active self-interference cancellation mechanisms in full-duplex wireless communication systems. In particular, we consider passive suppression due to antenna separation at the same node, and active cancellation in analog and/or digital domain. First, we show that the average amount of cancellation increases for active cancellation techniques as the received self-interference power increases. Our characterization of the average cancellation as a function of the self-interference power allows us to show that for a constant signal-to-interference ratio at the receiver antenna (before any active cancellation is applied), the rate of a full-duplex link increases as the self-interference power increases. Second, we show that applying digital cancellation after analog cancellation can sometimes increase the self-interference, and thus digital cancellation is more effective when applied selectively based on measured suppression values. Third, we complete our study of the impact of self-interference cancellation mechanisms by characterizing the probability distribution of the self-interference channel before and after cancellation.

1,398 citations

Journal ArticleDOI
TL;DR: A survey of existing solutions and standards is carried out, and design guidelines in architectures and protocols for mmWave communications are proposed, to facilitate the deployment of mmWave communication systems in the future 5G networks.
Abstract: With the explosive growth of mobile data demand, the fifth generation (5G) mobile network would exploit the enormous amount of spectrum in the millimeter wave (mmWave) bands to greatly increase communication capacity. There are fundamental differences between mmWave communications and existing other communication systems, in terms of high propagation loss, directivity, and sensitivity to blockage. These characteristics of mmWave communications pose several challenges to fully exploit the potential of mmWave communications, including integrated circuits and system design, interference management, spatial reuse, anti-blockage, and dynamics control. To address these challenges, we carry out a survey of existing solutions and standards, and propose design guidelines in architectures and protocols for mmWave communications. We also discuss the potential applications of mmWave communications in the 5G network, including the small cell access, the cellular access, and the wireless backhaul. Finally, we discuss relevant open research issues including the new physical layer technology, software-defined network architecture, measurements of network state information, efficient control mechanisms, and heterogeneous networking, which should be further investigated to facilitate the deployment of mmWave communication systems in the future 5G networks.

1,041 citations


Cites background or methods from "Practical, real-time, full duplex w..."

  • ...Miura and Bandai [127] proposed a full-duplex node architecture with directional transmission and omnidirectional reception based on the full-duplex architecture in [122]....

    [...]

  • ...received from a local transmitting antenna is usually much stronger than signal received from other nodes [122]....

    [...]

  • ...First, the current full-duplex systems have the bandwidth at the order of 40 MHz [122], and the practical implementation of the full-duplex systems for the mmWave communications with a bandwidth of several GHz should be investigated and demonstrated....

    [...]

  • ...[122] use the balanced/unbalanced (balun) transformer to support wideband and high power systems....

    [...]

References
More filters
Journal ArticleDOI
Emre Telatar1
01 Nov 1999
TL;DR: In this paper, the authors investigate the use of multiple transmitting and/or receiving antennas for single user communications over the additive Gaussian channel with and without fading, and derive formulas for the capacities and error exponents of such channels, and describe computational procedures to evaluate such formulas.
Abstract: We investigate the use of multiple transmitting and/or receiving antennas for single user communications over the additive Gaussian channel with and without fading. We derive formulas for the capacities and error exponents of such channels, and describe computational procedures to evaluate such formulas. We show that the potential gains of such multi-antenna systems over single-antenna systems is rather large under independenceassumptions for the fades and noises at different receiving antennas.

12,542 citations

Proceedings ArticleDOI
Jung-Il Choi1, Mayank Jain1, Kannan Srinivasan1, Phil Levis1, Sachin Katti1 
20 Sep 2010
TL;DR: In this paper, a single channel full-duplex wireless transceiver is proposed, which uses a combination of RF and baseband techniques to achieve FD with minimal effect on link reliability.
Abstract: This paper discusses the design of a single channel full-duplex wireless transceiver. The design uses a combination of RF and baseband techniques to achieve full-duplexing with minimal effect on link reliability. Experiments on real nodes show the full-duplex prototype achieves median performance that is within 8% of an ideal full-duplexing system. This paper presents Antenna Cancellation, a novel technique for self-interference cancellation. In conjunction with existing RF interference cancellation and digital baseband interference cancellation, antenna cancellation achieves the amount of self-interference cancellation required for full-duplex operation. The paper also discusses potential MAC and network gains with full-duplexing. It suggests ways in which a full-duplex system can solve some important problems with existing wireless systems including hidden terminals, loss of throughput due to congestion, and large end-to-end delays.

1,623 citations


"Practical, real-time, full duplex w..." refers background or methods in this paper

  • ...Motivated by the above limitations, recent work has proposed antenna placement techniques [4, 5]....

    [...]

  • ...Well known digital and analog techniques, even when combined together, are not sufficient to cancel self-interference sufficiently for full duplex [4]....

    [...]

  • ...propose an additional technique, called antenna cancellation [4]....

    [...]

  • ...Hence in our current implementation we have to make do with a component that provides an approximation, the QHx220 noise cancellation chip [11], which prior full duplex designs have used [4, 12]....

    [...]

  • ...The state of the art in full duplex operates on narrowband 5MHz signals with a transmit power of 0dBm (1mW) [4]....

    [...]

Proceedings ArticleDOI
01 Nov 2010
TL;DR: If the self-interference is cancelled in the analog domain before the interfering signal reaches the receiver front end, then the resulting full-duplex system can achieve rates higher than the rates achieved by a half-dulex system with identical analog resources.
Abstract: We study full-duplex wireless communication systems where same band simultaneous bidirectional communication is achieved via cancellation of the self-interfering signal Using off-the-shelf MIMO radios, we present experimental results that characterize the suppression performance of three self-interference cancellation mechanisms, which combine a different mix of analog and digital cancellation Our experimental results show that while the amount of self-interference increases linearly with the transmitted power, the self-interference can be sufficiently cancelled to make full-duplex wireless communication feasible in many cases Our experimental results further show that if the self-interference is cancelled in the analog domain before the interfering signal reaches the receiver front end, then the resulting full-duplex system can achieve rates higher than the rates achieved by a half-duplex system with identical analog resources

977 citations


"Practical, real-time, full duplex w..." refers background in this paper

  • ...Motivated by the above limitations, recent work has proposed antenna placement techniques [4, 5]....

    [...]

  • ...[5], which uses two separate TX chains....

    [...]

  • ...Antenna separation uses the fact that the distance between the transmit and receive antennas naturally reduces self-interference due to signal attenuation [5]....

    [...]

  • ...One approach to analog cancellation uses a second transmit chain to create an analog cancellation signal from a digital estimate of the self-interference [5], canceling ≈33dB of self-interference over a 625kHz bandwidth signal....

    [...]

  • ...The separation between the receive and transmit antennas attenuates the self-interference signal [5], but this antenna separation is not enough....

    [...]

Journal ArticleDOI
17 Aug 2008
TL;DR: This paper presents ZigZag, an 802.11 receiver design that combats hidden terminals, a new form of interference cancellation that exploits asynchrony across successive collisions in order to bootstrap its decoding.
Abstract: This paper presents ZigZag, an 802.11 receiver design that combats hidden terminals. ZigZag's core contribution is a new form of interference cancellation that exploits asynchrony across successive collisions. Specifically, 802.11 retransmissions, in the case of hidden terminals, cause successive collisions. These collisions have different interference-free stretches at their start, which ZigZag exploits to bootstrap its decoding.ZigZag makes no changes to the 802.11 MAC and introduces no overhead when there are no collisions. But, when senders collide, ZigZag attains the same throughput as if the colliding packets were a priori scheduled in separate time slots. We build a prototype of ZigZag in GNU Radio. In a testbed of 14 USRP nodes, ZigZag reduces the average packet loss rate at hidden terminals from 72.6% to about 0.7%.

671 citations


"Practical, real-time, full duplex w..." refers background in this paper

  • ...Implementing DC for a full duplex radio, however, is more challenging than other uses of digital cancellation, such as successive interference cancellation (SIC) [8] and ZigZag decoding [7]....

    [...]

  • ...Digital cancellation, while helpful, is by itself insufficient: current systems in the research literature cancel up to 20-25 dB [8, 7]....

    [...]

Proceedings ArticleDOI
11 May 2003
TL;DR: A codebook design method for quantized versions of maximum ratio transmission, equal gain transmission, and generalized selection diversity with maximum ratio combining at the receiver is presented and systems using the beamforming codebooks are shown to have a diversity order of the product of the number of transmit and thenumber of receive antennas.
Abstract: Multiple-input multiple-output (MIMO) wireless systems provides capacity much larger than that provided by traditional single-input single-output (SISO) wireless systems. Beamforming is a low complexity technique that increases the receive signal-to-noise ratio (SNR), however, it requires channel knowledge. Since in practice channel knowledge at the transmitter is difficult to realize, we propose a technique where the receiver designs the beamforming vector and sends it to the transmitter by transmitting a label in a finite set, or codebook, of beamforming vectors. A codebook design method for quantized versions of maximum ratio transmission, equal gain transmission, and generalized selection diversity with maximum ratio combining at the receiver is presented. The codebook design criterion exploits the quantization problem's relationship with Grassmannian line packing. Systems using the beamforming codebooks are shown to have a diversity order of the product of the number of transmit and the number of receive antennas. Monte Carlo simulations compare the performance of systems using this new codebook method with the performance of systems using previously proposed quantized and unquantized systems.

439 citations

Frequently Asked Questions (14)
Q1. What are the contributions in "Practical, real-time, full duplex wireless" ?

This paper presents a full duplex radio design using signal inversion and adaptive cancellation. This new design, unlike prior work, supports wideband and high power systems. This paper also presents a full duplex medium access control ( MAC ) design and evaluates it using a testbed of 5 prototype full duplex nodes. In practice, the authors find that the signal inversion technique alone can cancel at least 45dB across a 40MHz bandwidth. Further, combining signal inversion cancellation with cancellation in the digital domain can reduce self-interference by up to 73dB for a 10MHz OFDM signal. 

A balun cancellation-based full duplex radio, while an important step forward, leaves several open questions and new possibilities and future research. Exploring the effect on the TCP layer may be an interesting future work. 

Balun cancellation has no bandwidth constraint: it can in theory support arbitrary bandwidths and cancel arbitrarily high transmit powers. 

Transmission overlap depends on solving two technical challenges, minimizing secondary response latency and having transmission flexibility in preloaded packets. 

The QHx220 can cancel 20-25 dB for a 10MHz bandwidth signal but it introduces numerous complications, such as non-linearities and distortions, which complicate digital cancellation significantly. 

While an SIC implementation that recovers 80% of otherwise lost packets is a tremendous success, a full duplex radio that drops 20% of packets is barely usable. 

The basic approach is to estimate the attenuation and delay of the self-interference signal and match the inverse signal appropriately. 

Since the training symbols are defined in the frequency domain – each OFDM subband is narrow enough to have a flat frequency response – the radio estimates the frequency response of the self interference channel as a complex scalar value at each subcarrier. 

The key reason is that the balun circuit is not frequency flat, i.e., different parts of the band are inverted with different amplitudes. 

Digital cancellation, while helpful, is by itself insufficient: current systems in the research literature cancel up to 20-25 dB [8, 7]. 

The results in Figure 6 show that, if the phase and amplitude of the inverted signal are set correctly, balun cancellation can have impressive results across a wide bandwidth. 

Unlike prior designs, balun cancellation can work on wideband, high power signals, such that it is possible to build full duplex 802.11n devices. 

Unlike SIC or ZigZag, which use DC to recover packets which would have otherwise been lost, a full duplex radio uses DC to prevent the loss of packets which a half duplex radio could receive. 

Each node maintains per-destination transmission queues, and the hardware auto responder logic automatically picks the correct transmission queue to send a secondary packet from based on the header of the primary reception.