1Department of Electronic and Electrical Engineering, University College London, ... 4Aston Institute of Photonic Technologies, Aston University, Birmingham, ...
Tu2G.2.pdf
OFC 2015 © OSA 2015
Experimental Verification of Visible Light Communications based on Multi-Band CAP Modulation Paul Anthony Haigh1,2 , Petr Chvojka2 , Stanislav Zvanovec2 , Zabih Ghassemlooy3 , Son Thai Le4 , Thavamaran Kanesan5 , Elias Giacoumidis 4 , Nick J. Doran4 , Ioannis Papakonstantinou1 and Izzat Darwazeh1
'HSDUWPHQWRI(OHFWURQLFDQG(OHFWULFDO(QJLQHHULQJ8QLYHUVLW\&ROOHJH/RQGRQ/RQGRQ:&(%78. )DFXOW\RI(OHFWURPDJQHWLF)LHOGV&]HFK7HFKQLFDO8QLYHUVLW\LQ3UDJXH7HFKQLFND3UDKD&]HFK5HSXEOLF )DFXOW\RI(QJLQHHULQJDQG(QYLURQPHQW1RUWKXPEULD8QLYHUVLW\1HZFDVWOHXSRQ7\QH1(678. $VWRQ,QVWLWXWHRI3KRWRQLF7HFKQRORJLHV$VWRQ8QLYHUVLW\%LUPLQJKDP%(78. 7HOHNRP5HVHDUFK 'HYHORSPHQW 705 ' 70,QQRYDWLRQ&HQWUH&\EHUMD\D6HODQJRU0DOD\VLD $XWKRUHPDLODGGUHVV SDXODQWKRQ\KDLJK#LHHHRUJ
Abstract: A multi-band CAP system is experimentally demonstrated for the first time in VLC. We show that with an 8-CAP testbed spectral efficiencies (~4.75 b/s/Hz) at a realistic distance of 1 m can be reached. O CIS code s: (060.2605) Free-space optical communication; (170.4090) Modulation techniques
1. Introducti on Severe bandwidth limitation introduced by light-emitting diodes (LEDs) is the major bottleneck in high-capacity visible light communications (VLC) [1]. The most commonly proposed technique to improve the VLC data rate is the use of modulation formats that offer high spectral efficiency, such as orthogonal frequency-division multiplexing (OFDM) [2]. Recently we have been witnessing reduced levels of research activities in OFDM for VLC as the carrier-less amplitude and phase modulation (CAP) has been shown to outperform OFDM when using the same experimental link. In CAP, the carrier frequencies are generated using finite impulse response filters, which are costeffective and easily reconfigurable. This offers substantial reduction in computational complexity in comparison to OFDM systems requiring (inverse) fast Fourier transforms. However, VLC can be modelled as a 1st order low pass filter [2], where frequencies outside the 3 dB modulation bandwidth will experience higher attenuation. This is a problem for CAP systems, which require a flat-band frequency response. Hence, in [3] a multi-band CAP (P-CAP) system was proposed for a short-reach fibre link, where the bandwidth was divided into six bands, showing substantial improvements in chromatic dispersion tolerance and signal to noise ratio (SNR) performance in comparison to conventional CAP, as well as a significant reduction in sampling frequency. In this work, we show, for the first time, that P-CAP is an effective method for improving the bandwidth usage in VLC. We experimentally investigate P-CAP with an increasing number of sub-bands, P = {8, 4, 1} and demonstrate that transmission speeds (spectral efficiencies) of 30.88 (4.75), 23.65 (3.6), 9.04 (1.4) Mb/s (b/s/Hz) can be achieved using generic LEDs (with 4.5 MHz bandwidth) for P = {8, 4, 1}, respectively, at a realistic transmission distance of 1 m. 2. Test Setup The schematic block diagram of the proposed link is depicted in Fig. 1, incoporating a photograph of the experimental setup inset. The generation of conventional and P-CAP is well analyzed in the literature [3] so is not covered here due to space constraints. Digital signal processing (DSP) is performed in MATLAB while LabVIEW is used to control test and measurement instruments. The TTI TGA12104 arbitrary waveform generator (AWG) was selected due to it’s large memory (1 Mpoints) and high arbitrary clock frequency (100 MHz). The signal frequency is given as the ratio of the length of the data sequence to the desired transmission frequency, hence due to the high length of the transmission sequences, the transmission bandwidth is limited to 6.5 MHz, which is a restriction imposed by the AWG. The bandwidth of the LEDs used is 4.5 MHz, meaning an additional 2 MHz is required for communications and hence there is an out-of-band transmission, where high frequency attentuation can be expected. The transmission distance is ~1 m; which is substantially longer compared to the majority of experimental works in VLC reported in the literature [2]. The light is focused onto the photodetector (OSD15-5T) using a 25 mm biconvex lens with a focal distance of 25 mm. The signal is amplified using a differential transimpedance amplifier (AD8015) (not shown in Fig. 1) before digitization by an Agilent DSO9254A real-time oscilloscope, represented as a W = WV sampler in Fig. 1. Following signal capture, the signal processing is carried offline in MATLAB. BPSK is transmitted in the first instance, in order to gain an estimate of the SNR available in each sub-band [5]. Using this data (see Fig. 2(a)), higher order 0P -QAM constellations are assigned to each sub-band (Fig. 2(b)). Even though >105 bits were captured, the BER target is set at 10-3 in order to provide a small error allowance in consideration of the 7% and 20% forward error correction (FEC) BER limits of 3.8×10-3 and 2×10-2 , respectively.
978-1-55752-937-4/15/$31.00 ©2015 Optical Society of America
Tu2G.2.pdf
OFC 2015 © OSA 2015
Fig. 1 Schematic block diagram of the system under test including a photograph inset; it should be noted that the “Up” and “Down” blocks indicates up-sampling and down-sampling, respectively Several SNR thresholds dictate the maximum number of bits-per-symbol-per-sub-band, based on the theoretical QAM BER relationships (dashed lines and colour blocks in Fig. 2(a)). Fig. 2(a) and (b) shows that for P > 1 the number of bits-per-symbol can be optimized for the measured SNRs, while for P = 1, the threshold for 4-QAM is not met. Instead of selecting BPSK for the 1-CAP modulation format we decided that it was advantageous to persist with 4-QAM at the cost of a higher BER (~10-2 ), which is still beneath the 20% FEC limit.
Fig. 2 (a) Measured SNR from the BPSK tests for every m; and (b) number of bits-per-symbol for every m The P-CAP signals optimized with the appropriate number of bits-per-symbol-per-sub-band are subsequently loaded into the TTI TGA12104 high-speed data transmission occurs. The received signal is downsampled and demapped from the 0P -QAM constellation to give an estimate of the transmitted symbols. The estimates are compared with the actual transmitted symbols to provide a symbol-by-symbol BER performance characterisation for each value of P. 3. Results The aggregated BER performance for each value of P is shown in Fig. 3(a), where a BER of 10-3 is achieved as expected, indicating the correct functionality of the system for P > 1.
Fig. 3(a) Measured BER performance for every value of m; showing the BER for m =1 (~10 -2 ) and for m > 1 (~10 -3 ) as expected; and (b) measured EVM for each sub-band for every version of m-CAP tested; showing a clear improvement with increasing m It is clear that the worst performance occurs for the 1-CAP system, which offers a BER of ~10-2 , failing to meet the BER target due to selection of 4-QAM; but is comfortably within the 20% FEC limit as expected. Further insight is gained by analyzing the error vector magnitude (EVM)-per-sub-band as illustrated in Fig. 3(b), which does not include 1-CAP (EVM = 43.32%).
Tu2G.2.pdf
OFC 2015 © OSA 2015
A discussion of the individual P-CAP performance will commence, starting with 8-CAP. The received and normalized electrical spectra for 8-CAP, 4-CAP and 1-CAP are shown in Fig. 4(a), (b) and (c), respectively, with several sub-band constellations, shown inset, that are highlighted in their respective colours. The gross transmission speed that can be achieved using 8-CAP at a BER of 10-3 is 33.2 Mb/s. After the removal of the 7% FEC overhead, a net 30.88 Mb/s data rate is supported, corresponding to a spectral efficiency of 4.75 b/s/Hz; the highest reported in this work. To the best of the authors’ collective knowledge, this is the highest spectral efficiency reported in any VLC systems operating over a distance of ~ 1 m with no time or frequency domain equalizers. Other systems have reported higher spectral efficiencies; for example 6.25 b/s/Hz was demonstrated in [2], however the link distance was only 0.05 m, a reduction of 20-fold in comparison to this work, and a frequency domain equalizer was used. We anticipate with a smaller distance the spectral efficiency could be improved well beyond the reported 6.25 b/s/Hz, due to additional optical power available across all sub-bands. The net transmission speeds that can be supported using the remaining {4, 1}-CAP systems after overhead removal are reduced to 23.65 and 9.04 Mb/s, respectively, in comparison to the 30.88 Mb/s available in 8-CAP; a reduction of at least 7.23 Mb/s and at most 21.84 Mb/s. The measured spectral efficiencies are therefore reduced to 3.6 and 1.4 b/s/Hz at a BER of 10-3 , respectively for {4, 1}-CAP. Finally, we infer that the improvements in spectral efficiency are limited for values of P > 10 in this system with a marginal gain shown between P = 8 and 10, however, the results are not shown here due to space requirements, and are shown in our detailed discussion of this work in [4]. The results outlined in this paper are significant because they show the substantial improvement offered when using P-CAP, achieving high transmission speeds accessible without using any equalizers or further DSP techniques, as shown in previously reported works [2, 6]. For more detailed analysis of P-CAP systems readers are referred to [4].
Fig. 4 The measured electrical spectrum of: (a) 8-CAP, (b) 4-CAP and (c) 1-CAP showing increasing attenuation of the sub-bands with decreasing m; inset are the constellations for the highlighted sub-bands 4. Conclusion We have experimentally demonstrated a VLC P-CAP system, for the first time, showing a remarkable improvement in spectral efficiency over traditional CAP systems. The signal bandwidth was fixed at 6.5 MHz and the LED bandwidth was 4.5 MHz. The number of bands in the system was set to P = {8, 4, 1}. We observed that for a higher number of sub-bands, a higher throughput could be supported, up to ~31.5 Mb/s with a spectral efficiency of 4.75 b/s/Hz. This is the highest reported value in the literature for a transmission link span of 1 m, to the best of our knowledge. Reducing the number of sub-bands will lead to lower throughput and spectral efficiency due to the wider bandwidths of the sub-bands, which are more prone to high frequency attenuation in such low-pass VLC systems. 5. References [1] [2] [3] [4] [5] [6]
L. Grobe, A. Paraskevopoulos, J. Hilt, D. Schulz, F. Lassak, F. HartliebHWDO, "High-speed visible light communication systems," ,(((&RPPXQLFDWLRQV0DJD]LQHvol. 51, pp. 60-66, 2013. D. T sonev, H. Chun, S. Rajbhandari, J. McKendry, S. Videv, E. GuHWDO, "A 3-Gb/s Single-LED OFDM-based Wireless VLC Link Using a Gallium Nitride μLED," 3KRWRQLFV7HFKQRORJ\/HWWHUV,(((vol. PP, pp. 1-1, 2014. M. I. Olmedo, Z. T ianjian, J. B. Jensen, Z. Qiwen, X. Xiaogeng, S. PopovHWDO, "Multiband Carrierless Amplitude Phase Modulation for High Capacity Optical Data Links," /LJKWZDYH7HFKQRORJ\-RXUQDORIvol. 32, pp. 798-804, 2014. P. A. Haigh, A. Burton, K. Werfli, H. Le Minh, E. Bentley, P. ChvojkaHWDO, "A Multi-CAP Visible Light Communications System with 4.85 b/s/Hz Spectral Efficiency," ,(((-RXUQDORQ6HOHFWHG$UHDVLQ&RPPXQLFDWLRQVaccepted for publications, 2015. R. A. Shafik, S. Rahman, and R. Islam, "On the Extended Relationships Among EVM, BER and SNR as Performance Metrics," in (OHFWULFDODQG&RPSXWHU(QJLQHHULQJ,&(&( ,QWHUQDWLRQDO&RQIHUHQFHRQ, 2006, pp. 408-411. F. M. Wu, C. T . Lin, C. C. Wei, C. W. Chen, Z. Y. Chen, H. T. HuangHWDO, "Performance Comparison of OFDM Signal and CAP Signal Over High Capacity RGB-LED-Based WDM Visible Light Communication," ,(((3KRWRQLFV-RXUQDOvol. 5, pp. 79015077901507, 2013.
