Coexistence Between UWB and Narrowband Cellular Systems

Transcription

1 Coexistence Between UWB and Narrowband Cellular Systems Öffentliche Diskussionssitzung des Fachausschusses 72 der ITG UWB - Technologien und mögliche Anwendungen Kamp-Lintfort, 11 November 2004 Christian Müller, Martin Mittelbach {muellerc,mittelbach}@ifnettu-dresdende Chair for Communications Theory Communications Laboratory p1

2 Outline Problem Approach Model Development Physical Model Statistical Model Combination Parameter Values Simulation Results Conclusions C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 2

3 I Problem C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 3

4 Problem Objective: Investigate the impact of cumulative interference from multiple UWB systems to a narrowband victim receiver d 0 cellular VR r UWB Main questions: What is the maximum acceptable UWB device density? Do FCC / ETSI spectral masks for UWB emission limits provide sufficient protection for narrowband systems? C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 4

5 II Approach C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 5

6 Approach Scenario specification: Cellular victim receiver (downlink) Indoor environment 3D UWB device distribution Activity model for UWB devices Characteristics Environment Propagation law Free-space range (d o ) Path loss exponent (n) Device density and distribution (ρ) specification: UWB interferer Power spectral density (G p ) Activity factor (p) Victim receiver Bandwidth, center freq (B VR, f c ) Accepted link budget degrad (M UWB ) Receiver noise (N 0 ) interfer- Co-channel ence (I co ) C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 6

7 Approach Model development: Physical Model (system char) single UWB device interference multiple UWB device interference victim receiver degradation Statistical Model (UWB activity) single UWB device probability distribution multiple UWB device probability distribution Combination combine interference and activity derive a value for the maximum degradation C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 7

8 Approach Provide base for investigations (solving method): Analytical solution Numerical solution Simulation C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 8

9 III Model Development PhysicalModel C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 9

10 Physical Model Geometry: UWB d 0 VR R 0 R 0 : Grid spacing R 0 = ρ 1/3 UWB interference to victim receiver: I UWB = i P i, P i = f (G p,b VR,d 0,r,f c,n) P i : Single UWB device interference, r: Victim receiver UWB distance C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 10

11 Physical Model Link budget degradation due to UWB: M UWB = f (ÎUWB,I co,n 0 ) Maximum acceptable UWB interference: Î UWB = g (M UWB,I co,n 0 ) I co : N 0 : Co-channel interference Receiver noise To calculate maximum acceptable UWB device density we have to compare: I UWB Î UWB specified by UWB, environment characteristics and regulation specified by victim receiver characteristics C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 11

12 Model Development Overview Physical Model (system char) single UWB device interference multiple UWB device interference victim receiver degradation C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 12

13 III Model Development Statistical Model C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 13

14 Statistical Model Assumptions regarding statistics of UWB communication: UWB devices are switched on at a certain probability p This value is called the activity factor Independent UWB device activity is assumed Tasks to be done: Derive a random variable and probability function that models the activity of a single UWB device Combine multiple random variables to model the activity of multiple UWB devices Since independent UWB device activity is assumed, use statistical product approach Derive probability function that models the activity of multiple UWB devices C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 14

15 Model Development Overview Physical Model (system char) single UWB device interference multiple UWB device interference Statistical Model (UWB activity) single UWB device probability distribution multiple UWB device probability distribution victim receiver degradation C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 15

16 III Model Development Combination C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 16

17 Combination Combine normalised UWB interference and UWB activity: What do we have so far? Expression I UWB for the UWB interference, such that devices are assumed to be switched on A random variable and probability function modelling the UWB device activity What will be the result? Expression I UWB (ω) for the UWB interference, such that devices are only switched on at probability p Compare actual and maximum acceptable UWB interference: What do we have so far? Expression ÎUWB for the maximum acceptable UWB interference What will be the result? Outage probability: P(M UWB x db) = P(I UWB (ω) ÎUWB), x [0, ) C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 17

18 Model Development Overview Physical Model (system char) single UWB device interference multiple UWB device interference victim receiver degradation Statistical Model (UWB activity) single UWB device probability distribution multiple UWB device probability distribution Combination combine interference and activity derive a value for the maximum degradation C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 18

19 IV Parameter Values C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 19

20 Parameter Values Get emission limits for UWB communication systems (G p )from FCC and ETSI mask: GSM-900 GSM-1800 FCC FCC modified ETSI -40 EIRP EIRP [dbm/mhz] [dbm] UMTS frequency [GHz] UWB emission limits for indoor systems C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 20

21 Parameter Values Parameters for physical model - single UWB device interference: parameter symbol values GSM-900 GSM-1800 UMTS wavelength [m] λ a bandwidth [MHz] B VR PSD FCC [dbm/mhz] PSD ETSI [dbm/mhz] G p free-space range [m] d 0 10 b path loss exponent n 40 c a frequency band GHz assumed b indoor, d0 taken from: Multiple Access Communications Ltd: An Investigation into the Potential Impact of Ultra-Wideband Transmission Systems, Feb 2000 c NLOS, see also: UCAN coexistence study - presentation, UWB cluster meeting, Sept 2003 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 21

22 Parameter Values Victim receiver noise and co-channel interference: parameter symbol values [dbm] GSM-900 GSM-1800 UMTS d receiver noise N e e e co-chnl interference I co e e f d UMTS voice service assumed e details on parameters and calculation in whylesscom coex deliverable [D54b] f value taken from: Radiocommunications Agency: Impact of UWB on Third- Generation Telecommunications, 2003 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 22

23 Parameter Values Maximum acceptable UWB interference: M UWB [db] Î UWB [dbm] GSM-900 GSM-1800 UMTS 10 g g Activity factor: parameter symbol values activity factor p 001, 0025, 005 h g taken from: Ericsson: Generic power spectral density limits for a single UWB interferer, 2002 h typical values as in: SwissCom: Study of Interference effects of a UWB mass deployment on GSM systems, 2003 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 23

24 V Simulation Results C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 24

25 Simulation Results Example Relative frequency of UWB interference compared against maximum acceptable interference given by GSM-1800 and M UWB =1dB FCC mask ETSI mask 006 p = 005, rho = p = 001, rho = relative frequency I 0 (GSM-1800, FCC, 1dB) relative frequency I 0 (GSM-1800, ETSI, 1dB) normalized interference normalized interference Simulation parameters: d 0 =1m, n =4, N d = 25576, N =10 6 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 25

26 Simulation Results Example Relative frequency of UWB interference compared against maximum acceptable interference given by GSM-1800 and M UWB =1dB FCC mask ETSI mask 006 p = 005, rho = p = 001, rho = relative frequency I 0 (GSM-1800, FCC, 1dB) relative frequency I 0 (GSM-1800, ETSI, 1dB) normalized interference normalized interference Simulation parameters: d 0 =1m, n =4, N d = 25576, N =10 6 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 26

27 Simulation Results Example Relative frequency of UWB interference compared against maximum acceptable interference given by GSM-1800 and M UWB =1dB FCC mask ETSI mask 006 p = 005, rho = p = 001, rho = relative frequency I 0 (GSM-1800, FCC, 1dB) relative frequency I 0 (GSM-1800, ETSI, 1dB) normalized interference normalized interference Simulation parameters: d 0 =1m, n =4, N d = 25576, N =10 6 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 27

28 Simulation Results Example Outage probabilities given by GSM-1800 and M UWB =1dB and FCC mask ETSI mask outage probability P( M(UWB)>1dB ) p = 005 p = 0025 p = 001 outage probability P( M(UWB)>1dB ) p = 005 p = 0025 p = UWB device density rho [1/m 3 ] UWB device density rho [1/m 3 ] P(M UWB x db) = 001 p P(M UWB x db) = 001 p ρ [1/m 3 ] ρ [1/m 3 ] C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 28

29 Simulation Results Overview mask M UWB [db] p permitted UWB device density ρ [1/m 3 ] GSM-900 GSM-1800 UMTS FCC ETSI Permitted UWB device density for P(M UWB x db) = 001 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 29

30 Simulation Results Overview mask M UWB [db] p minimum permitted distance [m] GSM-900 GSM-1800 UMTS FCC ETSI Minimum permitted UWB to victim receiver distance for P(M UWB x db) = 001 C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 30

31 VI Conclusions C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 31

32 Conclusions mask M UWB [db] p protection level GSM-900 GSM-1800 UMTS FCC ETSI = 0025 = 0010 = 0050 = = 0025 = = 0010 = = 0050 = 0025 = 0010 = = 0050 = = 0025 = = 0010 = = insufficient protection = sufficient protection exorbitant protection C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 32

33 Thanksforyourattention *** Are there any questions? C Müller, M Mittelbach Kamp-Lintfort, 11 November 2004 p 33