Docstoc

Beihang-new

Document Sample
Beihang-new Powered By Docstoc
					Wireless Coexistence in Open Radio 
 Spectrum: Curses and Blessings

                     Guoliang Xing

                   Assistant Professor
     Department of Computer Science and Engineering
               Michigan State University
                      Outline
• Wireless Coexistence in Open Radio Spectrum
  – ZigBee link quality assurance [ICNP10, best paper award]
  – WiFi-assisted time sync [MobiCom10, RTSS11]
• Collaborative Sensing in Cyber-Physical Systems
  – Diffusion profiling using robotic sensors [IPSN12]
  – Volcano monitoring [RTSS10]
• Barcode Streaming for Smartphones [MobiSys 12] 
                A Wireless Era
• Today’s world is replete with wireless devices
  – 750 M laptops, 1 B smartphones, tablets, routers, 
    remotes, baby monitors….
• Radios on same freq may generate interference
• Frequency resources are getting scarce




                                                     3
       Crowded 2.4 GHz Spectrum

• 2.4-2.5 GHz band is unlicensed
  – Wi-Fi, Bluetooth, ZigBee
  – Cordless phones, baby monitors, wireless 
    headsets….


• Wi-Fi interference is a growing concern
  – 59 M Wi-Fi units in 2005, 409 M in 2009, 1 B in 
    2012 

                                                       4
                      ZigBee Technology
   • Low communication power (10~50 mw)
   • Application domains
       – Smart energy, healthcare IT, Industrial/home automation, remote 
         controls, game consoles….

       – Ex: >10 million smart meters installed in the US




Smart thermostat (HAI )   Smart electricity meter (Elster) Industrial sensor networks
                                                                                           5
                                                               (Intel fabrication plant)
    Co-existence of Wi-Fi and ZigBee
• How bad (quantitatively) is the interference?

• Do state-of-the-art link techniques suffice?
  – If not, how do we enable efficient co-existence?


• Can we take advantage of the interference?



                                                       6
      Empirical Study of Coexistence
                                                                 WiFi interferer:
• Change WiFi node location                                         802.11g

• Measure ZigBee sending rate and      Interference
  packet delivery ratio                     link
                                                          Data link


                                                      ZigBee sender and recver
                                                         TelosB with CC2420




                                                                                 7
                       WiFi Interferer Position
           WiFi Hidden Terminals

• Don’t trigger backoff at ZigBee sender
• Corrupt packets at ZigBee receiver




                                                 8
                      WiFi Interferer Position
           Wi-Fi Blind Terminals
• Wi-Fi Interference on both ZigBee sender 
  and receiver
• Severe packet loss on ZigBee link
• WiFi sending rate not significantly affected




                                                 9
            Why Blind Terminals ?
• Power asymmetry                                          ZigBee
                                                          tx range
• Heterogeneous PHY layers                ZigBee sender

  – WiFi only senses de-
    modulatable signals
                                               ZigBee recver

  – Energy-based sensing?
                                WiFi
                             interferer

                                                     WiFi tx range



                                                                     10
   White Space in Real-life WiFi 
             Traffic
• Arrivals of Wi-Fi  frames     white space: cluster gaps
                              that can be utilized by ZigBee




• Large amount of channel idle time




                                                               11
   Modeling WiFi White Space
• Length of white space follows iid Pareto distri.


                                               α = 1ms
                                         shorter intervals are
                                         not usable for ZigBee




• Implementation
   • Collect white space samples in a moving time window
   • Generate model by Maximum Likelihood Estimation

                                                           12
                 Basic Idea of WISE
• Sender splits ZigBee frame into sub-frames
• Fill the white space with sub-frames
• Receiver assembles sub-frames into frame
         WiFi frame cluster     ZigBee sub-frames


ZigBee


                   Time
                              sampling window

                                        ZigBee frame
                                           pending     13
            Frame Adaptation
• Collision probability



Sub-Frame    White space
                                   ZigBee data rate
   size         age
                                       250Kbps

• Sub-frame size optimization    Collision
                                Threshold


                                Maximum ZigBee
                                  frame size
                                                      14
             Experiment Setting
• ZigBee configuration
   •   TelosB with ZigBee-compliant CC2420 radios
   •   Good link performance without WiFi interference

• WiFi configuration
   •   802.11g netbooks with Atheros AR9285 chipset
   •   D-ITG for realistic traffic generation
• Baseline protocols
   •   B-MAC and Opportunistic transmission (OppTx)
• Evaluation metrics
   •   Modeling accuracy, sampling frequency, delivery ratio, 
       throughput, overhead
                                                                 15
Frame Delivery Ratio




      Unicast with 3 retx
                            16
                      Outline
• Wireless Coexistence in Open Radio Spectrum
  – ZigBee link quality assurance [ICNP10, best paper award]
  – WiFi-assisted time sync [MobiCom10, RTSS11]
• Collaborative Sensing in Cyber-Physical Systems
  – Diffusion profiling using robotic sensors [IPSN12]
  – Volcano monitoring [RTSS10]
• Barcode Streaming for Smartphones [MobiSys 12] 
    Clock Sync in Sensor Networks
• Fundamental service in sensor networks
   • A network-wide common notion of time
   • Essential for data ordering and processing
• On-board clock suffers significant drift
   • Drift rate of crystal oscillator in TelosB is 30-50 ppm
   • Frequent synchronization is needed across network
• Hardware-based solutions
   • GPS, WWVB
   • Cost, power consumption, poor coverage

                                                               18
                        Key Idea
ØWi-Fi access points broadcast periodic beacons
ØSense beacons using ZigBee radio
  •Sampling wireless signals via Received Signal Strength (RSS)
ØSynchronize according to extracted beacons
                                    Periodic beacon signal




                  TM




                                                                  19
Spatial Coverage of AP




Coverage of 5 APs on the third floor of
   Engineering Building @ MSU             20
      Temporal Stability of Beacons
• 4 laptops at different locations for 2 days
• Logging all beacon frames, traffic rate and etc.  




                                                       21
               Challenges
• How to identify Wi-Fi beacons?
  – Many data frames between two beacons
  – Beacon period may be unknown!
               Finding Needle in a Haystack
                ZigBee radio
                                       ZigBee Sensor


                    RSS Sampling &         Common 
                                                                  Beacon Detector
   WiFi                Shaping           Multiple Folding 
Access Point

                                       amplify periodic signals

                                                     threshold




                                 100                                            23
                   Evaluation
n 19 TelosB motes with TinyOS 2.1
nSync to production Wi-Fi in MSU Engineering building
n10 continuous days of evaluation  




                                                        21
                      Outline
• Wireless Coexistence in Open Radio Spectrum
  – ZigBee link quality assurance [ICNP10, best paper award]
  – WiFi-assisted time sync [MobiCom10, RTSS11]
• Collaborative Sensing in Cyber-Physical Systems
  – Diffusion profiling using robotic sensors [IPSN12]
  – Volcano monitoring [RTSS10]
• Barcode Streaming for Smartphones [MobiSys 12] 
Harmful Diffusion Processes




    Unocal oil spill               BP oil spill,           Waste Pollution
Santa Barbara, CA, 1969      Gulf of Mexico, 2010         UK, 2009, Reuters
http://en.wikipedia.org     http://en.wikipedia.org


• Diffusion profiling
   • source location, concentration, diffusion speed
   • high accuracy, short delay
• Physical uncertainties
     – temporal evolution, sensor biases, environmental noises
04/19/2012                   IPSN'12, Beijing, China                          26
Traditional Approaches

• Manual sampling
     – labor intensive
     – coarse spatiotemporal
       granularity

• Fixed buoyed sensors
     – expensive, limited coverage, poor adaptability

• Mobile sensing via AUVs and sea gliders
     – expensive (>$50K), bulky, heavy


04/19/2012              IPSN'12, Beijing, China         27
Aquatic Sensing via Robotic Fish




             Smart Microsystems Lab, MSU


• On-board sensing, control, and wireless comm.
• Low manufacturing cost: ~$200-$500
• Limited power supply and sensing capability

04/19/2012         IPSN'12, Beijing, China   28
Problem Statement

                                                       diffusion source

                                                       robotic sensors




•Maximize profiling accuracy w/ limited power supply
     •Collaborative sensing: source location, concentration, speed
     •Scheduling sensor movement to increase profiling accuracy

04/19/2012                  IPSN'12, Beijing, China                       29
Overview of Our Approach

• Maximum likelihood based estimation
• New estimation accuracy metric
    – Decouple sensors’ contributions
• New movement scheduling algorithm
    – Near-optimal dynamic programming
• Evaluation based on real data traces



04/19/2012           IPSN'12, Beijing, China   30
                      Outline
• Wireless Coexistence in Open Radio Spectrum
  – ZigBee link quality assurance [ICNP10, best paper award]
  – WiFi-assisted time sync [MobiCom10, RTSS11]
• Collaborative Sensing in Cyber-Physical Systems
  – Diffusion profiling using robotic sensors [IPSN12]
  – Volcano monitoring [RTSS10]
• Barcode Streaming for Smartphones [MobiSys 12] 
                                Volcano Hazards




Eruption in Chile, 6/4, 2011                                     Eruptions in Iceland 2010
$68 M instant damage, $2.4 B future relief.                      A week-long airspace closure
www.boston.com/bigpicture/2011/06/volcano_erupts_in_chile.html   [Wikipedia]


  • 7% world population live near active volcanoes
  • 20 - 30 explosive eruptions/year
                                                                                                32
                     Volcano Monitoring
• Seismic station
   – Expensive (~ $10K), bulky, difficult to 
     install, up to a dozen of nodes for most 
     active volcanoes!

• Data collection and retrieval
   – ~10G data in a month

• Processing
   – Detection, timing
   – 4D Tomography computation
       • Real-time, 3D fluid dynamics of a volcano 
         conduit system
   – Extremely computation-intensive 
            VolcanoSRI Project
• Large-scale, long-term deployment
  – 100~500 nodes/volcano, 1-year lifetime 
• Collaborative in-network processing
  – Detection, timing, localization
  – 4D tomography computation




              The tentative deployment map at Ecuador 
              (Photo credits: Prof. Jonathan Lees)
                 Approach Overview
                                               system decision

                                FFT
                                  ‘1’
seismic sensor
                                               sensor selection
                                        ‘0’    decision fusion

                      ‘1’
                     FFT
                                         FFT

• Select sensors with best signal qualities
   – FFT (computation-intensive)
• Local detection
• Decision fusion           avoid raw data transmission
                                                         35
   Sensing Fidelity Verification

             IOIO board
 Amplifier
                                       Seismometer
                                    Geospace Geophone 
                                      model GS-11D

              External     LG GT540
                GPS       Android 1.6

  GPS
antenna
                      Outline
• Wireless Coexistence in Open Radio Spectrum
  – ZigBee link quality assurance [ICNP10, best paper award]
  – WiFi-assisted time sync [MobiCom10, RTSS11]
• Collaborative Sensing in Cyber-Physical Systems
  – Diffusion profiling using robotic sensors [IPSN12]
  – Volcano monitoring
• Barcode Streaming for Smartphones [MobiSys 12] 
    Near Field Communication (NFC)

• Commonly used for Smart Payment




• Limits the communication to a short range (10cm)
• Only supported by a few smartphone platforms


                                                     38
                   COBRA System
• Real-time visible light communication (VLC)
system for off-the-shelf smartphones
   - Sender encodes info into color barcodes
   - Barcodes are streamed (15 fps) from screen to camera
   - Receiver decodes barcodes to get info




                    Streaming barcodes
                    btw screen and camera

          sender                            receiver
                                                            39
   System Overview




QR code
               Acknowledgement 
• Group members
  – Tian Hao (Ph.D, 2010-), Yu Wang (Ph.D, 2010-), Jun Huang (Ph.D, 2009-
    ), Ruogu Zhou (Ph.D, 2009-), Dennis Philips (Ph.D, 2009-), Jinzhu Chen 
    (Ph.D, 2010-), Mohammad-Mahdi Moazzami (Ph.D, 2011-), Fatme El-
    Moukaddem (Ph.D, co-supervised with Dr. Eric Torng), Rui Tan 
    (Postdoc)

• Research Sponsorship (~1.5 M USD since 2009)
  – NSF CDI, VolcanoSRI, 2011-2015 (in collaboration with WenZhan Song 
    @ GSU, Jonathan Lees@University of North Carolina, Chapel Hill)
  – NSF CAREER, performance-critical sensor networks, PI, 2010-2015.
  – NSF ECCS, aquatic sensor networks, PI, 2010-2013
  – NSF CNS, Interference in crowded spectrum, MSU PI, 2009-2012 (in 
    collaboration with Gang Zhou @ William & Mary)
  – Nokia University Cooperation Award


                                                                         41
            MSU CSE Ranking
• National Research Council's (NRC) 2011
  – R-ranking 10%-25%, S-ranking 8%-35% of 126
  – Overall 17%
•  Communications of the ACM
  – 17th of all US CSE graduate programs 
         Representative Publications
• Top conference publications since 2008
     – RTSS (8), MobiCom (2), MobiSys (2), SenSys (1), IPSN (1), MobiHoc (1),
       ICNP (2), Infocom (3), ICDCS (2), PerCom (1)

• Google Scholar: total # of citations since 2007: 2014, H-Index 20
•   J. Huang, G. Xing, G. Zhou, R. Zhou, Beyond Co-existence: Exploiting WiFi White Space for 
    ZigBee Performance Assurance, The 18th IEEE International Conference on Network 
    Protocols (ICNP), 2010, acceptance ratio: 31/170 = 18.2%, Best Paper Award (1 out of
    170 submissions).
•   R. Zhou, Y. Xiong, G. Xing, L. Sun, J. Ma, ZiFi: Wireless LAN Discovery via ZigBee 
    Interference Signatures, The 16th Annual International Conference on Mobile Computing 
    and Networking (MobiCom), acceptance ratio: 33/233=14.2%. 
•   T. Hao, R. Zhou, G. Xing, M. Mutka, WizSync: Exploiting Wi-Fi Infrastructure for Clock 
    Synchronization in Wireless Sensor Networks, IEEE Real-Time Systems Symposium (RTSS), 
    2011, acceptance ratio: 21%. 
•   S. Liu, G. Xing, H. Zhang, J. Wang, J. Huang, M. Sha, L. Huang, Passive Interference 
    Measurement in Wireless Sensor Networks, The 18th IEEE International Conference on 
    Network Protocols (ICNP), acceptance ratio: 31/170 = 18.2%, Best Paper Candidate (6
    out of 170 submissions).
•   R. Tan, G. Xing, J. Chen, W. Song, R. Huang, Quality-driven Volcanic Earthquake Detection 
    using Wireless Sensor Networks, The 31st IEEE Real-Time Systems Symposium (RTSS), 
    2010.
•   J. Chen, R. Tan, G. Xing, X. Wang, X. Fu, Fidelity-Aware Utilization Control for Cyber-
    Physical Surveillance Systems, The 31st IEEE Real-Time Systems Symposium (RTSS), 2010.
•   X. Xu, L. Gu, J. Wang, G. Xing, Negotiate Power and Performance in the Reality of RFID 
    Systems, The 8th Annual IEEE International Conference on Pervasive Computing and 
    Communications (PerCom), acceptance ratio: 27/227=12%, Best Paper Candidate (3 out
    of 227 submissions) .
                                                                                          43
    Challenge 1: Spatial Diversity




            Two earthquakes on Mt St Helens

• Complicated physical process
  – Highly dynamic magnitude
  – Dynamic source location
                                              44
   Challenge 2: Frequency Diversity




                  [1 Hz, 5 Hz]         [5 Hz, 10 Hz
Signal energy: X 10000            X 100

• Responsive to P-wave within [1 Hz, 10 Hz]
• Freq. spectrum changes with signal magnitude
                                                 45

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:7/2/2014
language:Latin
pages:45