Docstoc

Slide 1 - Clemson University

Document Sample
Slide 1 - Clemson University Powered By Docstoc
					Interference Effects of
Multi-User
Ultra-wideband Systems

               Anup Doshi
 Carnegie Mellon University   July 31, 2003
Outline
 Intro


 Models


 Observations


 Summary
        What is an Ultra-wideband Signal?
 Short impulses in                                                       Impulse Signal
                                                            1
        succession
                                                           0.5

                                                            0
 FCC Definition –
                                                          -0.5
    Bandwidth > 25% of
     center frequency                                       -1
                                                              0   200    400          600    800    1000
                          Power Spectrum Magnitude (dB)                          ns
                                                                                PSD
                                                            0

   1
                                                           -20
 0.5
                                                           -40
   0




                                                           -60
 -0.5




                                                           -80

                                                          -100
                                                              0   2000   4000         6000   8000   10000
                                                                                MHz
   Advantages of UWB
    Low power levels spread over large spectrum
             Operates below noise floor of narrowband devices
                    GPS
                          PCS                     802.11a
                                                            “Part 15 Limit”
-41 dBm/Mhz

                                                  UWB
                   1.6 1.9       3.1         5    Spectrum           10.6
                                Frequency (Ghz)                   Source: Intel



    Possibility of >500Mbps short range
Potential Applications are Numerous
 Personal Area Network                             UWB
              UWB
      Interconnect Computers, Devices, PDAs, Printers
      Entertainment...TV, Camcorder, DVD
        UWB

      Music…MP3, Audio Systems, etc
                                 LAN/WLAN


 Safety
      Through-wall Imaging
      Sensor Network
       UWB                                             UWB
 Lots of other exciting applications
                    Broadband




                                  Image Sources: Intel, AetherWire
Why Only Now?

    Started as impulse radar, 1960’s
       Primitive forms, simple communication

       Studied & used by military




    New technology enables digital comm., 1990’s
       Commercial applications seen by several companies

       1998 - Petitioned FCC to review potential uses

       2003 - FCC approves development of conservative

        applications
Problem…
 What happens when lots of UWB devices are
  transmitting in close proximity?
 Will the combined noise level be too much for a
  victim narrowband receiver?
 Existing studies claim minimal effects
      Done by various agencies and companies
 Those studies do not examine all cases…

           This is my job!
Constant-Distance Distribution
Multiple UWB devices located
three meters from a victim




                                 VICTIM
Characterizing the Transmitters
 Units turn on and off in a 2-state Markov Process
                         λ

              Unit                Unit
              Off                 On

                         µ
 Switching times are Exponential Random Variables
    Time until on ~ Exponential(λ) => mean 1/λ sec
    Time until off ~ Exponential(µ) => mean 1/µ sec


 Rho=ρ= λ/µ
Characterizing the Transmitters
 Total Number on modeled as a Markov Chain

       Nλ         (N-1)λ          …              λ


  0           1            2               N-1        N


        µ          2µ             …              Nµ

 Steady-state probabilities:
                         N  n
                   pn   
                         n  (1   ) N
                         
How Does the System Act Over Time?

    1

    2

    3

    4

    5

    6

    7

    8

    9

   10
        0   0.5   1   1.5   2   2.5   3   3.5   4

   λ =1, µ=2
How Does the System Act Over Time?

    1

    2

    3

    4

    5

    6

    7

    8

    9

   10
        0   0.5   1      1.5    2    2.5    3    3.5   4

   λ =1, µ=2          Total Number of Units On
Noise Level in Victim Receiver
 Each UWB signal modeled as White Noise
      Total Noise= N0+M(t)*N1



Ambient Noise Floor                     Power Received at
(=kTw)                                  Victim from UWB Signal


                Number of Transmitters On
                (Markov Chain)
                  Some Properties of This Model
                   Autocorrelation                                             Spectral Density
                       Rx(tau) = E(M(t)M(t+tau))                                   PSD = fft(Rx(t))
            2.5                                                    8


                                                                   7
             2
                                                                   6


            1.5                                                    5
Magnitude




                                                       Magnitude
                                                                   4

             1
                                                                   3


                                                                   2
            0.5

                                                                   1

             0                                                      0
             -5                   0                5               -100   -50             0           50   100
                              tau (sec)                                                  Hz
Probability of Error in Receiver
                            k                 2 Eb     
 On Average: Perr   P( M (t )  i ) *Q
                                                       
                                                        
                     i 1                 N 0  i * N1 
                       Average Probabilty of Error
          0.35


           0.3


          0.25
                         20 units

           0.2
   Perr




          0.15

                                            10 units
           0.1


          0.05


            0
             0   0.5    1            1.5        2      2.5   3
                                    rho (µ=1)
Other Ways to Describe Model
 Probability of Outage
    P(outage)=Probability( Perr > Pe* )
       Pe*=.1, .01


                        2 Eb        
                  
           Perr= Q                  
                                    
                   N 0  M (t ) N 1 

 Expected Time of Outage
    E(T10) = T1+aN,1E(T20)
       E(T20)=T2+aN,2E(T10)+bN,2E(T30)

       …

       E(TN0)=TN+E(T(N-1)0)
                     P(outage), Expected Time of Outage

                                    P(outage)                                                    E(outage)
             1                                                                    12

            0.9
                  12 units                                                        10
            0.8
                             10 units
            0.7
                                                                                  8
            0.6
P(outage)




                                                                        t (sec)
            0.5                                                                   6
                                                                                                                    12 units
            0.4
                                                                                  4
            0.3

            0.2
                                                                                  2                                            10 units
            0.1

             0                                                                    0
              0       0.5    1           1.5        2   2.5         3              0   0.5   1      1.5         2          2.5             3
                                        rho (µ=1)                                                  rho (µ=10)
                                                        3m radius                                                              3m radius
Random-Distance Distribution
UWB devices distributed
uniformly in a circular area
around victim




                               VICTIM
Properties of Random-Distance Model
 Moved to a computer simulation


 Experimentally calculated:
      P(outage),
      Expected Time of Outage,
      Max and mean power levels over time

 Done on Matlab – Monte Carlo simulation
Example Simulation Run
                        Interference Level
       -34

       -36

       -38

       -40

       -42
 dBm




       -44

       -46

       -48

       -50

       -52
          0   0.5   1   1.5     2      2.5   3   3.5   4
                              time
Example Simulation Run
                           Interference Level
        -30

        -40

        -50

        -60

        -70
 dBm




        -80

        -90

       -100

       -110

       -120
           0   1   2   3       4          5     6   7   8   9
                                   time
Example Simulation Run
                   Interference Level
        -30

        -40

        -50

        -60

        -70
 dBm




        -80

        -90

       -100

       -110

       -120
           0   5           10           15   20
                         time
Example Simulation Run
                        Interference Level
        -30

        -40

        -50

        -60

        -70
 dBm




        -80

        -90

       -100

       -110

       -120
           0   5   10   15      20      25   30   35   40
                              time
                      P(outage) & Expected Time of Outage

                                  Probability of Outage                                               Expected Time of Outage
                                                                                           9
                Random-distance
           1                                                                               8
                radius=3m

                                                                                           7
          0.8
                                                                                           6




                                                                              time (sec)
                             Constant-distance
Poutage




          0.6                                                                              5
                             radius=3m
                                                                                           4          Random-distance
                                                                                                      radius=3m
          0.4
                                                                                           3

                                                                                           2
          0.2                                                                                                               Constant-distance
                                                                                                                            radius=3m
                                                                                           1

           0                                                                               0
            0         0.5         1        1.5            2   2.5         3                 0   0.5   1         1.5           2         2.5              3
                                           rho (µ=1)           10 units
                                                                                                               rho (µ=10)                     10 units
Max/Mean Power Levels
              Average Max and Mean Power Levels over 500 simulations
      -20

             Average Max Power
      -30


      -40

                                             Allowed Interference Level
      -50
dBm




                 Average Mean Power
      -60


      -70


      -80


      -90
         0       5      10     15      20       25      30     35         40
                       Number of Transmitters in 10m radius
Observations
 Multiple transmitters will cause major
  problems in worst cases

 Such situations may soon arise in real-life
  situations

 Important to consider every possible case in
  testing

                     Broadband
Future Work
 Need to consider many more variables
      Receiver type
      Frequency, PRR
      Different distributions
 Once 802 Standard comes out, incorporate into
  model
      Possibly Multi-Band OFDM (TI, Intel)
      Possibly Dual-Band (Time Domain, Motorola)
Summary
 Characterized an aggregate of UWB transmitters


 Realized various methods of measuring effect on
  victim receiver

 Concluded that as number of UWB transmitters
  increase, performance of victim receiver
  attenuates
Acknowledgments
 Prof Baum


 Prof Noneaker, Prof Xu


 ECE Faculty and Grads


 NSF

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:2
posted:2/27/2012
language:
pages:29