Spectrum Efficient Coexistence with Primary Users George Nychis Ranveer Chandra Thomas Moscibroda Ivan Tashev Peter Steenkiste Carnegie ID: 384435
Download Presentation The PPT/PDF document "Reclaiming the White Spaces:" is the property of its rightful owner. Permission is granted to download and print the materials on this web site for personal, non-commercial use only, and to display it on your personal computer provided you do not modify the materials and that you retain all copyright notices contained in the materials. By downloading content from our website, you accept the terms of this agreement.
Slide1
Reclaiming the White Spaces:Spectrum Efficient Coexistencewith Primary UsersGeorge Nychis†, Ranveer Chandra§, Thomas Moscibroda★, Ivan Tashev§, Peter Steenkiste† †Carnegie Mellon University, §Microsoft Research, ★Microsoft Research Asia
1Slide2
dbmFrequency-60-100
“White spaces”
470 MHz
700
MHz
What are White Spaces?
0
MHz
7000
MHz
TV
ISM (Wi-Fi)
700
470
2400
5180
2500
5300
are
Unoccupied
TV Channels
White Spaces
54-90
170-216
2
Wireless
Mic
More Spectrum
Longer Range
9 Orthogonal Channels (150Mbps per
chan
)
at least
3 - 4x of Wi-Fi
}
Potential Applications
Rural wireless broadband
City-wide mesh
……..
……..
Proliferation of Wireless & Mobile:
ISM Band is Insufficient to Meet DemandSlide3
White Space AvailabilitySpectrum availability is critical to adoption and goalsSpectrum is most critical in populated areasMeasure spectrum availability in top 30 U.S. cities [1]3[1] Geo-location database: http://whitespaces.msresearch.us/“[to] make a significant amount of spectrum available for new and innovative products and services” – FCC
08-260
53% of cities cannot
support single 802.11
channel in the white spaces
Number of
Analog TV
Broadcasts
Decreasing
Over Time
s
ingle
c
hannel
2 Reserved Channels
Losing a
Significant
Amount of
White Space
to
Mic
Rules
Reclaiming the White Spaces
Goal: Rescue White Space by Enabling
Mic
CoexistenceSlide4
OutlineBackground on mic signalsData transmission impact on mic audioCritical insight on coexistenceSEISMIC: Spectrum Efficient Interference-Free System for MICsRegains spectrum with zero audible interferenceEvaluation of effectiveness and efficiency4Slide5
Analog Microphone Background5Mic ReceiverWireless Mic
One-way communication
: receiver never transmits
Mic
always
transmitting
(even idle)
Signal Components
:
FM Modulated Audio Signal
Used By
Mic
Receiver to Detect Low Signal and MuteSlide6
OutlineBackground on mic signalsData transmission impact on mic audioCritical insight on coexistenceSEISMIC: Spectrum Efficient Interference-Free System for MICsRegains spectrum with zero audible interferenceEvaluation of effectiveness and efficiency6Slide7
How Do You Coexist to Reclaim Spectrum?First in-depth analysis RF interference on mic audio quality3 Dimensions: time, frequency, and powerStudy Impact on 6 Mics: Audio Technica, Sennheiser (3), Shure, and Electro-Voice7
First in-depth analysis RF interference on
mic
audio quality
3 Dimensions:
time
,
frequency, and powerSlide8
Experimental Setup
2. MIC Recording
to Computer
1. PC Output
to Speakers
PESQ
worldwide audio evaluation standard
Compare recording to original
0 (total disruption)
1 (perfect)
White Space Device
Mic
Receiver
MicSlide9
Interference in FrequencyFix power and duration, vary frequency (25KHz steps)92. Suppression Req. Varies by Mic
1. Avoid Disruption, Still Use 97% of ChannelSlide10
Interference in PowerFix freq. and duration, vary power of WSD (2dB steps)10
WSN
m
(dB)
WSN
s
(dB)
Record value of
WSN
m
& WSN
s
when PESQ=1
No audio disruption when interference is
just below the squelch tones! (WSN
s
= 1dB)
Despite 25dB of noise!
“FM receivers exhibit a `capture effect’ in which they respond to only the strongest signal received on a frequency and reject any weaker interfering signals.”
- FCC (First Order: 04-113)Slide11
Implications of Interference StudyCoexistence in the same channel is possible!Great! Just suppress bandwidth required at center frequency11Not that simple…Slide12
The Challenge of SuppressionWeaker Mic More Suppression
12
Stronger WSD
More Suppression
`Perfect Suppression’
would have sharp edge
Actual Suppression
Leaks Power in to the BandSlide13
13Two components needed at WSD
to suppress properly:
Information Needed to Suppress
White Space Device
(WSD)
Mic
Receiver
Wireless
Mic
2.
Mic
Signal Power
at
Mic
Receiver
1. WSD Interference Power
at the
Mic
ReceiverSlide14
No feedback on either required componentsWithout feedback, the system is open-loopMust suppress “worst-case” to be conservative14Lack of Information at WSDUnfortunately, worst-case is vacation! (6MHz)
Given the
open-loop
state of the system (info available)…
… the FCC made the right decision by requiring vacation.Slide15
Need a Closed-Loop to Avoid Vacation15White Space Device)Mic ReceiverMic
Measurement
Feedback
Analysis
Adaptation
SEISMICSlide16
OutlineBackground on mic signalsData transmission impact on mic audioCritical insight on coexistenceSEISMIC: Spectrum Efficient Interference-Free System for MICsRegains spectrum with zero audible interferenceEvaluation of effectiveness and efficiency16Slide17
SEISMIC System OverviewImplements closed-loop design to avoid vacation:Measurement: MicProtector measures key components at receiverFeedback: “Strobe” signal to notify WSD of impending disruptionAnalysis / Adaptation: SEISMIC protocol to adapt frequency17
White Space Device
Mic
Receiver
Mic
MicProtector
FeedbackSlide18
Measurement: MicProtectorKey measurement needed: squelch & interference powerSquelch Measurement: estimated in the mic’s bandInterference Measurement: estimated in control bands18FrequencyAmplitudeControl
Band
Control
Band
25KHz
25KHz
Interference
Level
Power in
Mic
Band ShiftsSlide19
Feedback: StrobesMicProtector needs to convey information to WSDStrobe is similar to Morse-codes and on/off-keying (OOK)Allows us to convey necessary information without complex protocol (e.g., 802.11)Strobe signals convey: (1) impending disruption, (2) required bandwidth, (3) center frequency19Frequency
AmplitudeSlide20
Adaptation / Analysis: SEISMIC ProtocolWSD: sends probe packets with increasing power (exploit capture)MicProtector: notifies of impending disruption, using strobes20FrequencyAmplitudeProtection Threshold
Control
Band
Control
Band
25KHz
25KHz
Interference
LevelSlide21
WSD: sends probe packets with increasing power (exploit capture)MicProtector: notifies of impending disruption, using strobes21ProbeStrobe
Pkts
:
Time
WSD
MicProt
.
Suppressed Frequency (KHz)
Increase in Power
MicProtector
Strobes the WSD
for interference near threshold
50
100
100
150
200
250
250
250
Convergence
To Coexistence
Adaptation:
SEISMIC ProtocolSlide22
Summary of SEISMIC DesignMicProtector components:detection of impending audio disruptionsfeedback using strobes to WSD enables closed-loopNote: can be built directly in to future receiversWSD and MicProtector engage in SEISMIC protocolconverge to optimal suppression around micFormalization shows correctness with multiple WSDWSDs will converge and never interfere, in paper 22
S
E
IS
M
I
CSlide23
OutlineBackground on mic signalsData transmission impact on mic audioCritical insight on coexistenceSEISMIC: Spectrum Efficient Interference-Free System for MICsRegains spectrum with zero audible interferenceEvaluation of effectiveness and efficiency23Slide24
Evaluating SEISMICFull MicProtector and SEISMIC WSD prototypescustom USRP2 builds with UHF front ends Evaluation on several coexistence points:Effectiveness: ability to avoid audio disruptionEfficiency: ability to enable high spectrum re-useSingle microphone scenarioMany microphone scenario24Slide25
Effectiveness of SEISMIC’s CoexistenceChallenge: Low-power & mobile microphoneSEISMIC WSD must never interfere despite mic signal fluctuations25Mobility createsquick fluctuations
WSD continually
and quickly
adapts to avoid
audio disruption!
WSD vacates
channel when
mic
power is
low to be safeSlide26
Efficiency of SEISMIC’s CoexistenceEvaluate efficiency under two mobile mic scenariosFar mic (low/moderate signal) .. close WSD (high interference)Mod. mic (moderate signal) .. nearby WSD (moderate interference26Mic Distance (ft)
WSD
Distance (
ft
)
1.
Far
(50-70)
Close
(5)
2.
Moderate
(10-30)
Moderate
(15)
Common scenario:
95% of time >5.7MHz
Challenging scenario:
70% of time >5.2MHz
Low
(-95
-75)
Moderate
(-65)
Mic Signal (dBm)
High (-50-30)
Moderate (-70)
WSD Signal (
dBm)Slide27
SEISMIC Efficiency with Many MicsObtain real mic freq placement from coordinatorsModel components: 1) Mic signals, 2) WSDs & interferenceGiven components, how much spectrum does WSD X have?27
Event
Mics
2008 NBA All-Star-Game
191
2010 BCS Championship
108
2010
WPC Conference
77
SEISMIC-enabled
Mic
System
SEISMIC-enabled WSD
Mic
signals generated using mobile
mic
measurements
Adaptrum
WSD
freq
suppression
w/ leakage used for WSDsSlide28
SEISMIC Spectrum Efficiency28
On average
, SEISMIC-enabled WSDs:
SEISMIC
has 22x, 3.6x, and 1.6x availability compared to
channel vacation
SEISMIC
near
perfect suppression
,
suppresses only when necessary
(closed-loop)
With 150MHz:
can support 7 150Mbps channels! (2.5x 2.4GHz, with 4x range)Slide29
Summary29Spectrum availability is critical to white spacesCoexistence possible between WSD and micsClosed-loop solution required to avoid vacationSEISMIC enables disruption-free coexistenceUp to 95% of the spectrum can be regainedDemoed to FCC chairman, mic manufacturers, mic coordinators, audio communitySlide30
Deployment ChallengeIf a mic receiver does not have a MicProtector…3030
Time
Power
WSD View
Mic
System View
(No
MicProtector
)
Power
Frequency
No strobe is detected, WSD ramps up…
… eventually begins disruptingSlide31
Fortunately, Can Partially DeployCurrent ruling: all mics must report to DBAugment database to include SEISMIC capabilityOnly use channel if all mics have MicProtectorMore SEISMIC receivers better spectrum efficiency31
System
ID
Location
…
SEISMIC-Enabled
aa:bb:cc
35.669743,
139.762595
…
No
xx:yy:zz
35.669743,
139.762595
…
YesSlide32
RF Interference Impact StudyFirst in-depth analysis on mic audio qualityOpposed to traditional “RF-interference” measurementStudy Impact on 6 Mics: Audio Technica, Sennheiser (3), Shure, and Electro-VoiceUnderstand impact of 3 key components:Time: can short transmissions avoid audio disruption?Frequency: what if only part of the channel used?Power: will low power transmissions avoid disruption?32
Recorded
Mic
Audio
Transmit 16μs long data packets (1/100
th
of 802.11 packet)
spaced by 500ms in timeSlide33
Effectiveness of SEISMIC’s CoexistenceChallenge: Low-power & mobile microphoneSEISMIC WSD must never interfere despite mic signal fluctuations33Mobility createsquick fluctuations
-98dBm USRP2 noise floor,
10dB Protection Threshold,
MicProtector
considers
mic
low power @ -88dBm
WSD continually
and quickly
adapts to avoid
audio disruption!
SEISMIC signals
WSD to vacate
channel when
mic
power is
low to be safeSlide34
Efficiency of SEISMIC’s CoexistenceEvaluate efficiency under two mobile mic scenariosFar mic (low/moderate signal) .. close WSD (high interference)Mod. mic (moderate signal) .. nearby WSD (moderate interference34Mic Distance (ft)
WSD
Distance (
ft
)
1.
Far
(50-70)
Close
(5)
2.
Moderate
(10-30)
Moderate
(15)
Common scenario:95% of time >5.7MHz
Challenging scenario:
70% of time >5.2MHz
Low
(-95
-75)
Moderate
(-65)Mic Signal (
dBm)High (-50
-30)Moderate (-70)
WSD Signal (dBm)Slide35
Closing the LoopTo avoid vacation we need a closed-loop system35Mic ReceiverWhite Space Device
Measurement
Feedback
Adaptation
Analysis
SEISMICSlide36
White Space AvailabilitySpectrum availability is critical to adoption and goalsSpectrum is most critical in populated areasMeasure spectrum availability in top 30 U.S. cities [1]36[1] Geo-location database: http://whitespaces.msresearch.us/Unfortunately:1. Availability is low
2
.
Mic
coexistence
rules lower availability
(2 channels reserved)
39% of cities have zero
white space available!
“[to] make a significant amount of spectrum available for new and innovative products and services” –
FCC
08-260
3.
Mics
can still operate
outside of reserved channels,
driving
availbility
even lower
53% of cities cannot
support single 802.11
channel in the white spacesSlide37
White Space AvailabilitySpectrum availability is critical to adoption and goalsSpectrum is most critical in populated areasMeasure spectrum availability in top 30 U.S. cities [1]37[1] Geo-location database: http://whitespaces.msresearch.us/Key Points:1. Availability is low
2
. Efficiency of the
spectrum use is
critical
“[to] make a significant amount of spectrum available for new and innovative products and services” –
FCC
08-260
53% of cities cannot
support single 802.11
channel in the white spacesSlide38
White Space AvailabilitySpectrum availability is critical to adoption and goalsSpectrum is most critical in populated areasMeasure spectrum availability in top 30 U.S. cities [1]38[1] Geo-location database: http://whitespaces.msresearch.us/Unfortunately:1. Availability is low
2
.
Mic
coexistence
rules lower availability
(2 channels reserved)
39% of cities have zero
white space available!
“[to] make a significant amount of spectrum available for new and innovative products and services” –
FCC
08-260
3.
Mics
can still operate
outside of reserved channels,
driving
availbility
even lower
53% of cities cannot
support single 802.11
channel in the white spacesSlide39
FCC Ruling: Unable to use channel with an active mic2 channels reserved for wireless microphonesReclaiming the White Spaces39single channelReclaim 95%SpectrumAround
Mic
Remove
Need For
2 Reserved
Channels
Goal: Reclaim White Space by Enabling
Mic
CoexistenceSlide40
Measurement: MicProtectorImplements three components:Interference Detection: estimated in control bandsInterference Protection: monitors squelch & interferenceImpending Interference Notification: strobe signals 40FrequencyAmplitudeProtection Threshold
Strobe Signals
Control
Band
Control
Band
25KHz
25KHz
Interference
Level