The Measurement and Characterisation of Ultra Wide-Band (UWB) Intentionally Radiated Signals

Transcription

1 The Measurement and Characterisation of Ultra Wide-Band (UWB) Intentionally Radiated Signals Rafael Cepeda Toshiba Research Europe Ltd University of Bristol November 2007

2 Structure of Presentation Introduction to UWB Measuring Equipment Channel Models and Examples Alternative/Complementary Methods Conclusions 2

3 Introduction to Ultra Wide-Band (UWB) 3

4 Definition of UWB UWB intentional emissions must occupy a bandwidth of at least 500 MHz and/or more than 25% fractional bandwidth ( ) f b PSD - Power Spectral Density Narrowband UWB f b = f h f c f l 25% -10dB -10dB fl fc f h > 500 MHz 4

5 Regulations Different regions, different UWB spectral masks DAA Europe* USA Regulations DAA EIRP max: db/mhz Bandwidth: ~ GHz DAA: Detect And Avoid * Provisional FCC: Federal Communications Commission Japan* EIRP: Equivalent Isotropically Radiated Power 5

6 Measuring Equipment 6

7 Types of UWB Sounders + 7

8 Characteristics of Sounders Pico-second pulse generator Pulse generated then captured by DSO (Digital Storage Oscilloscope) Off-the-shelf equipment Separated transceivers Only amplitude information Difficult to remove noise from measurements PN-sequence correlator PN-sequence generated then correlated at the receiver Length of PN-sequence determines the total measuring time Good noise rejection Difficult to process long PN-sequences Vector Network Analyser Frequency carriers are generated and received sequentially Whole dynamic range available per frequency carrier Time consuming Frequency spectrum not measured at once 8

9 UWB Channel Sounding System at TRL Transmit Antenna Positioners Controller for positioners Master controller Receive Antenna Channel sounder Bandpass filters Calibration equipment LNA Power amplifiers Spectrum analyser LNA: Low Noise Amplifier 9

10 System Specifications Sounder type: PN correlation (4095 chips) Bandwidth: ~ 7 GHz (3.5 GHz 10.5 GHz) Channel snapshots: ~100 per second Maximum channel length: 589 ns Transmitted Power: ~ 30dBm Spatial channels: 8 (2x4) Antennas: Bi-conical UWB sounder (MEDAV) 10

11 Bi-conical Antenna Radiation & VSWR Bi-conical antenna (IRK) Radiation at 6.4 GHz Protective casing Antenna VSWR Radiation at 6.4 GHz VSWR: Voltage Standing Wave Ratio 11

12 Antenna Gain & Co-polar Power vs. Frequency Antenna Gain Measured by the two-antenna method* Co-polar power contribution Calculated from 3D radiation patterns * C. A. Balanis. Antenna Theory - Analysis and Design. John Wiley & Sons,

13 Post-processing of Recorded Channel Data START Measured propagation signal Cross correlation Subtract cross-talk Cross-talk measurement Cross correlation Time alignment System response Cross correlation Subtract cross-talk Filter system response END CIR: Channel Impulse Response CIR Known PN sequence Threshold data AGC gain & attenuation compensation Time domain Frequency domain AGC: Automatic Gain Control 13

14 Comparison of VNA & TD* Sounder Measurements Anechoic chamber Rx antenna Tx antenna Metallic objects Servo positioner Channel sounding equipment * Time Domain 14

15 Channel Models and Measurement Example 15

16 UWB Channel Models UWB Channel is formed from clustered rays that follow a double log-normal distribution Γ Overall Envelope Amplitude γ Cluster Envelope Arriving rays nγ+mγ Time IEEE a Channel models (CM): CM1: LOS channel 0-4m CM3: NLOS 4-10m CM2: NLOS 0-4m CM4: fits a 25ns RMS delay spread (extreme NLOS) IEEE: Institute of Electrical and Electronic Engineers LOS: Line-of-Sight / NLOS: non LOS 16

17 Indoor LOS Channel Sounding 3.35m kx0 K0 ky0 4.15m 2 A TV B 1.82m Kitchen 3 Fireplace Living room m ly0 L0 lx0 lx0 3.63m PC Front room fy0 Tx Hall fx0 F0 1.3 m Rx Ø 0.1 m Up hx0 hx0 hy0 H0 1.30m Transmitter location Receiver location 2 m 17

18 Evidence of New Type of Channel Statistics Model: PL( f) PL f δ Pathloss Frequency f 2δ Frequency path-loss exponent Extreme exponents can arise! Shift: reflection, absorption? Number of occurrences Width: diffraction? Slope exponent [δ] Magnitude [db] Magnitude [db] Frequency [GHz] Frequency [GHz] 18

19 3.02m Results per Link Location Transmitter location Receiver location Number of occurrences Slope exponent [δ] 4.15m Number of occurrences A TV B 1.82m Number of occurrences Number of occurrences Slope exponent [δ] ly0 3 Living room 1 4 Fireplace 5 Number of occurrences Slope exponent [δ] L0 lx Slope exponent [δ] 3.63m Slope exponent [δ] 19

20 Results and Comparison with Other Measurements Frequency Path Loss Exponent Links Tx Rx ( δ ± σ ) 1 A 0.55 ± B 0.39 ± B 0.64 ± B 1.09 ± B 0.58 ± m reference 0 ± 0.02 All IRs 0.66 ± 0.55 Sanity check results: RMS Delay Spread (ns) Links Tx Rx ( τ ± σ) RMS 1 A 6.60 ± B 5.05 ± B 6.26 ± B 8.29 ± B 8.04 ± m reference 3.65 ± 0.02 All IRs 6.85 ± 1.62 [Malik: 2006] 0.51 ± 0.24 (horizontal) & 0.60 ± 0.20 (vertical) [Chong: 2005] 0.62 ±

21 NLOS Frequency Dependent Path-Loss (4-10 GHz) Y X 10 0 Performance with κ = 0-2, TFI (spread over 1.5 GHz) and 3/4-rate code Performance of UWB systems using 528 MHz bandwidth varies by 3 db 10-1 Using more bandwidth will increase this variation PER 10-2 no freq. losses δ = 0 δ = 0.5 δ = 1 δ = 1.5 δ = SNR (db) 21

22 Ray Tracing (deterministic) Tools Characterisation of the radio channel is essential for UWB system design, as pathloss is critical Antenna radiation patterns measured in anechoic chamber Modelling Modern house 22

23 Summary and Conclusions Brief introduction to UWB and measuring equipment; State-of-the-art equipment; Two contrasting sounding techniques; Channel models and measurement campaign; Alternative techniques to physical measurements. 23