Open Access
Add to BibSonomyAbstract
An experimental test-bed using a narrow-pulsed ultraviolet (UV) laser and high-bandwidth photomultiplier tube was set up to characterize pulse broadening effects in short-range non-line-of-sight (NLOS) scattering communication channels. Pulse broadening is reported as a function of the transmitter elevation angle, transmitter beam angle, receiver elevation angle, receiver field-of-view, and transmitter-receiver distance. The results provide insight into the channel bandwidth and achievable communication data rate.
©2010 Optical Society of America
1. Introduction
Atmospheric scattering of ultraviolet (UV) radiation by abundant molecules and aerosols provides a major mechanism for establishing a non-line-of-sight (NLOS) communication link. A series of relevant UV communication studies in the solar blind band (wavelength 200nm-280nm) has been conducted since the 1960s. These include channel characterization to system implementation and networking. An early survey on Navy communication applications and transceiver design was reported in [1,2]. Pioneering work focused on NLOS UV link characterization using a xenon flashtube as the light source and a photomultiplier tube (PMT) as the receiver [3]. In addition, an elegant analytical channel response model was developed to describe the temporal characteristics of scattered radiation in the middle UV [4]. The model was further extended to examine angular spectra and path loss [5,6]. Around the same time, a point-to-point NLOS UV communication system at a modulation rate of 40 kHz was experimentally demonstrated based on an isotropic radiating mercury arc lamp [7]; see also [8]. An improved UV local area network test-bed covering a 1km range and using a collimated mercury-xenon lamp was constructed with a modulation rate up to 400 kHz [9]. UV pulsed laser communication systems operating with rates of few hundred hertz were also reported [10,11].
Early long-range communication systems were based on bulky and power-hungry light sources such as flashtubes, lamps, or lasers. Recent advances in low-cost, small-size, low-power, high-reliability, and high-bandwidth deep UV light-emitting-diodes (LEDs) [12–14] and avalanche photodiodes (APDs) [15,16] have motivated further work on short-range communications, imaging, and sensor networks [17–22]. In particular, for LED-based outdoor UV communications, the authors and their colleagues have reported experimental performance results including receiver bit-error-rates with different transmitter-receiver geometries [23,24]. Other results include a simplified analytical single scattering path loss model [25], a semi-analytical multiple scattering path loss model [26], and Monte Carlo simulation multiple scattering path loss and impulse response models [27].
In this paper, we present results from several experiments incorporating different optical pointing and receiver field of view (FOV). Using short laser pulses and a high-speed PMT receiver, we measured the impulse response. The results depict pulse broadening as a function of transmitter (Tx) and receiver (Rx) elevation angles, optical source beam angle, receiver FOV, and Tx-Rx separation distance. These measurements provide a means to assess channel bandwidth and to infer achievable signaling rates, which are limited by inter-symbol interference (ISI) due to pulse broadening, as a function of the system parameters.
2. Channel impulse response test-bed
Our solar blind UV communication impulse response test-bed is depicted in Fig. 1 . The critical elements include a high power short-time pulsed UV source, a high sensitivity PMT detector, and a solar-blind filter with good out-of-band rejection.
The transmitter was a compact Q-switched fourth harmonic Nd:YAG laser at 266 nm triggered by a rectangular pulse at 10 Hz from a signal generator, producing a corresponding laser pulse train at 10 Hz. Each pulse had width of (3-5) ns and energy up to (3-5) mJ, and can be attenuated manually by a factor of 103 with an integrated attenuator. The laser head was mounted on a rotation stage with precise motorized angular control. A synchronization signal from the signal generator was output to the oscilloscope at the receiver through a cable. On top of the laser output window, an optical beam expander and a UV focus lens were mounted to adjust the beam angle. A set of beam angles was obtained and measured off-line using a set of lenses of different focal lengths.
For the optical source, there is no commercial laser diode in the solar blind band. A commercial UV LED typically delivers 1mw optical power, with LED arrays up to 50mw [12]. Given that the path loss at 100 m range could be more than 100dB [24,26], an LED provides insufficient power for reliable impulse response measurements. Moreover, current UV LED modulation bandwidths are only about 50 MHz and thus time domain resolution is more than 20 ns. Consequently, we employed a Q-switched solid state UV laser. A drawback is that the laser could not be modulated at continuously varying frequencies. Therefore, the frequency domain channel sounding technique successfully applied for indoor infrared channel testing is not applicable [28–30], and we used time domain waveform recording techniques.
At the receiver, we considered both APD and PMT detectors. The detector was followed by a customized high gain (34 dB) preamplifier module at 1.5 GHz. A 3 GHz oscilloscope was used to view and record the receiver output waveform. We first tested both detectors with LOS operation at negligible range. Results are shown in Fig. 2 . As we can see from Fig. 2a, the full-width half-maximum (FWHM) pulse width using the APD is less than 5 ns. The ripple in the tail was mainly due to electric discharge in the circuit. However, while the APD has a fast response time of 1 ns, it’s relatively small active area limits gain to only about 102-104. Consequently, the APD is not suitable for many NLOS measurement scenarios of interest due to inadequate weak signal response.
Figure 2b shows the measured response with a PMT with LOS at negligible range. While the PMT gain is very high (105-107), its multiple electrode structure tends to spread the response time. The fastest commercial UV PMT available at the time of the experiment was a Hamamatsu PMT whose response time was about 6 ns with an active diameter of 8mm. The FWHM pulse width is 6.2 ns, and this represents the minimum time resolution in our channel response measurements due to non-ideal devices in the system, including the laser, detector and amplifier circuit. However, as we show, the channel response typically spans tens to thousands of nano-seconds. So, we regard the measurement as the system impulse response and ignore the residual error. From Fig. 2, note that the system response waveform is close to Gaussian.
All subsequent results reported are based on a Hamamatsu PMT module H10304 with an integrated high voltage circuit and active diameter 8 mm, gain 2.3×106, and a response time of 6 ns. A focusing lens and a 1-inch solar blind UV filter were mounted before the PMT. The Rx FOV was easily changed by using a set of lenses of different focal lengths, as in the Tx beam angle control. A mechanical module incorporated a rotation stage to achieve high-resolution zenith angular control.
Referring to Fig. 1 and similarly as defined in [26], we denote the Tx beam full-width divergence by ϕ1, the FOV angle by ϕ2, the Tx elevation angle by θ1, Rx elevation angle by θ2, and the Tx and Rx baseline separation distance by r. In the following, we present our test results, linking these key system parameters in different Tx and Rx geometry with pulse broadening effects in terms of the waveforms and their FWHM widths.
3. Experimental results
A series of measurements were conducted including varying the Tx and Rx elevation angles from 0° to 90°, and changing the Tx beam angle, Rx FOV, and baseline distance up to 100 m. The experimental site was an outdoor open field in clear weather. The laser pulse rate was set at 10 Hz, and the output energy after the optical beam shaping system was 3 mJ. Beam divergence was less than 3 mrad, and line width was 1 cm−1. The 270 nm solar blind filter had FWHM bandwidth of 15 nm. The peak transmission was 10.4% at 271 nm, and the out-of-band transmission was less than 10−8 at 290 nm, 10−10 at 390 nm, and 10−11 at 305 nm - 750 nm. The spectral mismatch between the laser and the filter was found to be less than 30%. The PMT had a circular sensing window with a diameter of 8 mm, and its dark current was 4 nA. The transmitter’s beam angle was controlled at 3 mrad, 12°, and 25°. The receiver’s effective FOV without a focus lens was 30°. It became 13°, 9.2°, 4.58°, and 2° when using different focus lenses in front of the PMT.
3.1 Random variation across measurements
Each measured impulse response exhibits random variation including the changing atmosphere as well as the detector response. We average over realizations to yield the average response for a fixed set of system parameters.
Figure 3 shows the normalized waveforms after averaging 1, 5, 10, and 20 realizations in the same geometry, respectively. Note the significant spikes before averaging due to the discrete photon arrivals. The Tx and Rx elevation angles were 40° and 90°, beam angle was 3 mrad, the receiver’s effective FOV was 30°, with a baseline of 100 m. These results are typical, indicating the measurement variation around the mean behavior. In the following all results are based on averaging over 50 realizations.
3.2 Impulse response waveform shape
While the laser output pulse follows a Gaussian shape, the received waveform varies with different NLOS geometries. To gain some insight, Fig. 4 shows typical average impulse responses under different Tx and Rx elevation angles, progressing from small angle 10° to large angle 90°. The resulting FWHM pulse width varies from 20 ns to 1500 ns. Note the faster rising edge, followed by slower decay. The decay increases as the elevation angle increases, due to the larger spread in propagation times for scattered photons. With a smaller elevation angle (lower to the horizon), the decay becomes significantly sharper and the pulse width is decreased. Note also that, at higher elevation angles (approaching vertical pointing), the responses have more variations from measurement to measurement, and the averaged response continues to exhibit significant spread, even after averaging 50 measurements.
3.3 Comparison of different pulse width calculation methods
There are (at least) three widely used definitions to obtain a pulse width from a given waveform. First, FWHM is often adopted for a laser pulse. Second, the 3dB-equivalent width (or just “3dB width”), equal to the inverse of the 3dB bandwidth of the waveform, is also considered for wireless communication. The third is root mean square (rms) pulse spread σ, which has been used to describe an indoor infrared diffused channel response h(t) [28–30], defined as
In addition, for comparison, we also define the full pulse width at 5% maximum because a long tail with very small amplitude may deliver limited information. The FWHM and 5% pulse width can be directly read from a waveform plot. The rms width can be calculated from stored data samples according to Eq. (1). The 3dB width is obtained by taking the fast Fourier transform (FFT) of the waveform to obtain the 3dB bandwidth in the frequency domain. Table 1 compares these widths for fixed Tx beam angle 3 mrad, Rx FOV 30°, and separation distance 100 m.
Table 1. Comparison of four different pulse widths
We can see the following approximate relations among them: (a) the rms width is the smallest; (b) the FWHM width is about 3 times the rms width; (c) the 3dB width tracks and is roughly equal to the 5% width; and (d) the 3dB width and 5% width both are about 6 times the rms width. In the frequency domain, the 3dB bandwidths for the four pointing cases are (58.62, 5.70, 2.34, 0.46) MHz, respectively. Clearly, a small pointing angle leads to a large channel bandwidth. Due to the reciprocal relation between bandwidth and waveform width, we will focus on time domain waveform width in this paper. From these results, FWHM seems to be a good representative choice for pulse width, and we utilize it in the following.
3.4 Pulse broadening with transmitter and receiver elevation angles
We fixed the distance at 100 m, Tx beam angle 3 mrad, and Rx FOV 30°, while varying the Tx elevation angle from 10° to 90° and Rx elevation angle from 0° to 90°, each with a step size of 10°. For each pair of pointing angles, we recorded fifty impulse response waveforms and calculated the FWHM width of their average.
Figure 5 shows pulse width versus Tx elevation (apex) angle parameterized by different Rx elevation angles. We exclude the case of Tx 0° to avoid Tx LOS pointing at the Rx, considering NLOS geometry only. For a fixed Rx angle, the pulse width always increases with the Tx angle. Intuitively, the difference between the longest and shortest paths traversed by scattered photons increases, and thus the pulse broadens significantly. The rate of broadening increase depends on the Rx angle. For a small Rx angle, the pulse width increases slowly, while the rate of broadening increases dramatically for a large Rx angle. For example, in the case of Rx 10°, the pulse width increases from about 10 ns at Tx 10° to about 87 ns at Tx 90°. When the Rx increases to 90°, pulse width increases from 35 ns to 1100 ns. That implies the NLOS UV communication system bandwidth, approximated as the inverse of the pulse width, drops roughly from 29 Mbps to 999 Kbps. If we fix the Tx elevation angle, the pulse width increases with Rx elevation angle. For example, when the Tx angle is 90°, the pulse width increases from 87 ns to 1100 ns as the Rx elevation angle increases from 10° to 90°.
In order to better demonstrate how pulse width increases with Rx elevation angle for a fixed Tx elevation angle, the same data are plotted in Fig. 6 , now with pulse width versus Rx elevation angle and curves parameterized by Tx angle. The trend appears similar to Fig. 5 for the same pair of exchanged Tx/Rx angle values.
3.5 Pulse broadening with Tx beam angle
The emitted laser beam has a divergence angle of 3 mrad, or 0.176°. A beam expander and lens were used to reshape the laser beam, and two additional Tx beam angles (12° and 25°) were obtained by using two lenses with different focal lengths. Three sets of Tx ad Rx elevation angles (40°,40°), (40°,60°) (40°,90°) were considered, Rx FOV was 30° and the distance was 100 m. The FWHM results are presented in Table 2 . We observe only a weak dependence on the beam angles tested. As the beam angle increases, the pulse FWHM variation is typically less than 5%. This observation motivates examination of other types of pulse widths that can better capture the waveform tail effects. For example, let’s consider the 5% width. While the FWHM is relatively insensitive to beam angle variation, the 5% width shows a clear dependence. For fixed Tx and Rx angles, the 5% width typically increases monotonically with the beam angle, as observed in the last column of this table. The tail behavior may produce ISI, and may be a function of multiple scattering. Thus, in the communications context, the inverse of the pulse width is only a general indicator of channel bandwidth, and a careful study of ISI and channel capacity should incorporate the pulse waveform.
Table 2. Pulse width with 3 Tx beam angles
3.6 Pulse broadening with Rx field-of-view
Next we consider variation of the receiver FOV. A smaller FOV may generally lead to fewer received photons, and a corresponding reduction in signal strength. However, one could make a tradeoff between communication distance and achievable rate by controlling the FOV.
By mounting different lenses before the PMT, we obtained 5 FOVs: 30°, 13°, 9.2°, 4.58° and 2°. We conducted similar experiments to examine the FOV effect as we did for the Tx beam angle, while keeping other parameters fixed with Tx beam angle 3 mrad and a distance of 100 m.
Five pairs of Tx and Rx elevation angles were tested for each FOV, namely (40°,0°), (40°,30°), (40°,40°), (40°,60°), and (40°,90°). The corresponding pulse width results versus Rx FOV are presented in Fig. 7 . The pulse width increases linearly with the FOV for fixed elevation angles. A larger Rx angle induces a higher pulse broadening rate. The pulse width rises from 20 ns to 263 ns as the FOV increases from 2° to 30°. A smaller FOV simultaneously yields lower signal levels and smaller pulse bandwidth. Thus, if a reduced FOV is tolerable with regard to signal level, then a higher bandwidth channel is possible.
3.7 Pulse broadening with baseline distance
Experiments were also conducted to measure the pulse width changes with baseline distance. Three pairs of Tx and Rx elevation angles (40°,40°), (40°,60°), (40°,90°) were considered. The distance varied from 10 m to 100 m with a 10 m interval. The Tx beam angle was 3 mrad, and FOV 30°. The pulse width results are shown as a function of distance in Fig. 8 . With increasing distance, the pulse width roughly increases linearly. The larger the Rx angle is, the higher the rate of increase. When the distance increases from 10 m to 100 m, the pulse width increases from 10 ns to 87 ns at Rx angle of 40°. As Rx angle increases to 90°, the pulse width changes from 21 ns to 263 ns.
4. Conclusions and future work
We presented a variety of experimental pulse width measurements for short-range NLOS UV communication channels based on a pulsed solid state UV laser and fast response PMT. Pulse width dependence was studied for different system parameters, including Tx and Rx elevation angles, Tx beam angle, Rx FOV, and baseline distance. We found the pulse width was weakly dependent on Tx beam angle variations, but changed almost linearly with other parameters. These findings provide first-hand experimental channel information and are valuable for communication system design.
Further work based on measured waveforms will develop an empirical channel pulse broadening model, and compare with both Monte Carlo simulation and parametric model results [27]. Moreover, we will conduct channel tests for both path loss and pulse broadening when Tx beam axis and Rx FOV axis are non-coplanar, all of which can follow our approaches for coplanar geometry in this paper and [24,26]. Subsequently, we will use those results to guide our analytical modeling in non-coplanar geometry.
Acknowledgements
The authors would like to thank Kaiyun Cui, Haipeng Ding, Qunfeng He, Yiyang Li, and Leijie Wang for their invaluable help with experiments and discussions. This work was supported in part by the Army Research Office under Grants W911NF-08-1-0163 and W911NF-09-1-0293, and the Army Research Laboratory under the Collaborative Technology Alliance Program, Cooperative Agreement DAAD19-01-2-0011.
References and links
1. G. L. Harvey, “A survey of ultraviolet communication systems,” Naval Research Laboratory Technical Report, Washington D.C., March 13, 1964.
2. D. M. Junge, “Non-line-of-sight electro-optic laser communications in the middle ultraviolet,” M.S. Thesis, Naval Postgraduate School, Monterey, CA, December 1977.
3. D. E. Sunstein, “A scatter communications link at ultraviolet frequencies,” B.S. Thesis, MIT, Cambridge, MA, 1968.
4. D. M. Reilly, “Atmospheric optical communications in the middle ultraviolet,” M.S. Thesis, MIT, Cambridge, MA, 1976.
5. D. M. Reilly and C. Warde, “Temporal characteristics of single-scatter radiation,” J. Opt. Soc. Am. A 69(3), 464–470 (1979). [CrossRef]
6. M. R. Luettgen, J. H. Shapiro, and D. M. Reilly, “Non-line-of-sight single-scatter propagation model,” J. Opt. Soc. Am. A 8(12), 1964–1972 (1991). [CrossRef]
7. E. S. Fishburne, M. E. Neer, and G. Sandri, “Voice communication via scattered ultraviolet radiation,” final report of Aeronautical Research Associates of Princeton, Inc., NJ, February 1976.
8. W. S. Ross and R. S. Kennedy, “An investigation of atmospheric optically scattered non-line-of-sight communication links,” Army Research Office Project Report, Research Triangle Park, NC, January 1980.
9. J. J. Puschell and R. Bayse, “High data rate ultraviolet communication systems for the tactical battlefield,” Proc. of Tactical Communications Conf., 253–267 (1990).
10. B. Charles, B. Hughes, A. Erickson, J. Wilkins, and E. Teppo, “An ultraviolet laser based communication system for short range tactical applications,” Proc. SPIE 2115, 79–86 (1994). [CrossRef]
11. R. D. Shute, “Electrodeless ultraviolet communications system,” IEEE Aerosp. Electron. Syst. Mag. 10(11), 2–7 (1995). [CrossRef]
12. Sensor Electronic Technology Inc, http://www.s-et.com/.
13. M. Shatalov, J. Zhang, A. S. Chitnis, V. Adivarahan, J. Yang, G. Simin, and M. A. Khan, “Deep ultraviolet light-emitting diodes using quaternary AlInGaN multiple quantum wells,” IEEE J. Sel. Top. Quantum Electron. 8(2), 302–309 (2002). [CrossRef]
14. V. Adivarahan, Q. Fareed, S. Srivastava, T. Katona, M. Gaevski, and A. Khan, “Robust 285 nm Deep UV Light Emitting Diodes over Metal Organic Hydride Vapor Phase Epitaxially Grown AlN/Sapphire Templates,” Jpn. J. Appl. Phys. 46(23), 537–539 (2007). [CrossRef]
15. X. Bai, D. Mcintosh, H. Liu, and J. C. Campbell, “Ultraviolet single photon detection with Geiger-mode 4H-SiC avalanche photodiodes,” IEEE Photon. Technol. Lett. 19(22), 1822–1824 (2007). [CrossRef]
16. S. C. Shen, Y. Zhang, D. Yoo, J. B. Limb, J. H. Ryou, P. D. Yoder, and R. D. Dupuis, “Performance of deep ultraviolet GaN avalanche photodiodes grown by MOCVD,” IEEE Photon. Technol. Lett. 19(21), 1744–1746 (2007). [CrossRef]
17. D. Kedar and S. Arnon, “Non-line-of-sight optical wireless sensor network operating in multiscattering channel,” Appl. Opt. 45(33), 8454–8461 (2006). [CrossRef] [PubMed]
18. S. Arnon and D. Kedar, “Non-line-of-sight underwater optical wireless communication network,” J. Opt. Soc. Am. A 26(3), 530–539 (2009). [CrossRef]
19. D. Kedard and S. Arnon, “Subsea ultraviolet solar-blind broadband free-space optics communication,” Opt. Engin. 48, 046001 1–7 (2009).
20. C. Lavigne, G. Durand, and A. Roblin, “Ultraviolet light propagation under low visibility atmospheric conditions and its application to aircraft landing aid,” Appl. Opt. 45(36), 9140–9150 (2006). [CrossRef] [PubMed]
21. C. Lavigne, G. Durand, and A. Roblin, “Simulation comparison of aircraft landing performance in foggy conditions aided by different UV sensors,” Appl. Opt. 48(12), 2203–2213 (2009). [CrossRef] [PubMed]
22. G. A. Shaw, A. M. Siegel, and J. Model, “Extending the range and performance of non-line-of-sight ultraviolet communication links,” Proc. SPIE 62310C, 1–12 (2006).
23. Z. Xu and B. M. Sadler, “Ultraviolet communications: potential and state-of-the-art,” IEEE Commun. Mag. 46(5), 67–73 (2008). [CrossRef]
24. G. Chen, F. Abou-Galala, Z. Xu, and B. M. Sadler, “Experimental evaluation of LED-based solar blind NLOS communication links,” Opt. Express 16(19), 15059–15068 (2008). [CrossRef] [PubMed]
25. Z. Xu, H. Ding, B. M. Sadler, and G. Chen, “Analytical performance study of solar blind non-line-of-sight ultraviolet short-range communication links,” Opt. Lett. 33(16), 1860–1862 (2008). [CrossRef] [PubMed]
26. G. Chen, Z. Xu, H. Ding, and B. M. Sadler, “Path loss modeling and performance trade-off study for short-range non-line-of-sight ultraviolet communications,” Opt. Express 17(5), 3929–3940 (2009). [CrossRef] [PubMed]
27. H. Ding, G. Chen, A. Majumdar, B. M. Sadler, and Z. Xu, “Modeling of non-line-of-sight ultraviolet scattering channels for communication,” IEEE J. Sel. Areas Comm. 27(9), 1535–1544 (2009). [CrossRef]
28. M. R. Pakravan, M. Kavehrad, and H. Hashemi, “Indoor wireless infrared channel characterization by measurements,” IEEE Trans. Vehicular Technol. 50(4), 1053–1073 (2001). [CrossRef]
29. J. M. Kahn and J. R. Barry, “Wireless infrared communications,” Proc. IEEE 85(2), 265–298 (1997). [CrossRef]
30. J. M. Kahn, W. J. Krause, and J. B. Carruthers, “Experimental characterization of non-directed indoor infrared channels,” IEEE Trans. Commun. 43(2), 1613–1623 (1995). [CrossRef]
