White-Light LED Based Visible Light
Communication System
Volume 8, Number 6, December 2016
Chin-Wei Hsu
Chi-Wai Chow
I-Cheng Lu
Yen-Liang Liu
Chien-Hung Yeh
Yang Liu
DOI: 10.1109/JPHOT.2016.2633395
1943-0655 © 2016 IEEE
High Speed Imaging 3
×
3 MIMO
Phosphor White-Light LED Based Visible
Light Communication System
Chin-Wei Hsu,1Chi-Wai Chow,1I-Cheng Lu,2Yen-Liang Liu,2 Chien-Hung Yeh,3and Yang Liu4
1Department of Photonics and Institute of Electro-Optical Engineering, National Chiao Tung University, Hsinchu 300, Taiwan
2Industrial Technology Research Institute, Hsinchu 310, Taiwan
3Department of Photonics, Feng Chia University, Seatwen, Taichung 40724, Taiwan 4Philips Electronics Ltd. Hong Kong
DOI:10.1109/JPHOT.2016.2633395
1943-0655C 2016 IEEE. Translations and content mining are permitted for academic research only. Personal use is also permitted, but republication/redistribution requires IEEE permission. See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
Manuscript received October 27, 2016; revised November 16, 2016; accepted November 22, 2016. Date of publication December 7, 2016; date of current version December 15, 2016. This work was supported by the Ministry of Science and Technology, Taiwan, under Grant MOST-104-2628-E-009-011-MY3, Aim for the Top University Plan, Taiwan. Corresponding author: C.-W. Chow (e-mail: [email protected]).
Abstract: The phosphor light-emitting diode (LED) has been widely used not only for indoor illumination but for the optical transmitter (Tx) as well in visible light communication (VLC). However, the transmission data rate is significantly limited by the modulation bandwidth of LED because of the low-speed response of phosphor. In order to enhance the data rate of the VLC system, we demonstrate a 3×3 imaging multiple-input multiple-output (MIMO) VLC system. Besides, the pre-equalizer circuit is used to extend the bandwidth of the LED Tx. Orthogonal frequency division multiplexing (OFDM) with bit-loading algorithm is used. By using the proposed scheme, the original 1 MHz bandwidth commercially available phosphor white light LED can achieve 1 Gbps data rate over 1-m free-space transmission distance.
Index Terms:Light-emitting diode (LED), multiple-input multiple-output (MIMO), visible light communication (VLC).
1. Introduction
Nowadays, the light-emitting diode (LED) has been widely used for illumination because of its advantages such as low cost, long lifetime, and conservation of energy. In addition, it has been used as the light source for the visible light communication (VLC) [1]. Unlike laser based VLC systems [2]–[5], the LED based VLC can combine the lighting and communication at the same time. As the primary function of the LED is for lighting; hence little extra cost and energy is needed to provide the communication purpose. Previously, we have demonstrated the white light LED based VLC can also provide high precision indoor positioning [6] and light encryption [7]. In the LED based VLC system, the data rate is limited by the low modulation bandwidth (< 3 MHz) of phosphor white LED. Therefore, several methods have been proposed to increase the data rates. In [8], the use of an equalizer and blue-filter was reported to achieve a 100 Mbit/s non-return-to-zero (NRZ) over 0.1 m. Besides, advanced modulations, such as orthogonal frequency division multiplexing (OFDM) can provide a high spectrally efficiency, and [9] reported using OFDM, blue-filter and avalanche photodiode (APD) to achieve a 513 Mbit/s VLC over 0.3 m. To further increase
Fig. 1. Configuration of 3×3 MIMO VLC system. PD: photodiode.
the data rates, multiple-input multiple-output (MIMO) technique has been proposed and studied [10], [11]. In the non-imaging MIMO, separate focusing lens is required in front of each photodiode (PD). Non-imaging 4×4 MIMO VLC of 50 Mbit/s over 2 m was demonstrated [12]. The dedicated focusing lens for each PD made the size of VLC receiver (Rx) very big. In [13], use of an imaging 4×9 MIMO VLC of 1 Gbit/s over 1 m using an imaging lens, blue filter, and a 3×3 PD array is demonstrated; however, a complicated 3×3 PD array was required to achieve 1 Gbit/s. In [14], a 4×4 imaging MIMO achieving an aggregate data rate of 920 Mbit/s is reported; however, special
μLEDs were required, and this system may not good for lighting. In [15], another imaging MIMO imaging Rx is reported; however, only simulations were provided.
In this work, we propose and demonstrate a high speed 3×3 imaging MIMO VLC system using commercial phosphorescent white LEDs and low-cost PIN PDs. Only one focusing lens is needed at the Rx, and therefore, the size of Rx can be more compact than non-imaging MIMO system. Besides, commercially available white LEDs are used; hence the proposed MIMO system can perform lighting and communication simultaneously. By combining the equalizer circuit and OFDM bit-loading, the total data rate of the system can achieve 1 Gbit/s over 1 m free space transmission.
2. Operation Principle of 3
×
3 MIMO VLC System
Fig. 1 shows the configuration of 3×3 MIMO VLC system. Here, LEDs are used as transmitters (Txs) and the transmitted signal will be received by PDs. However, each PD will not only receive the light signal from one transmitter but the interference light from other channels as well. In addition, these interferences will distort the target signal so that the performance of that channel will deteriorate. Such multi-channel communication system can be expressed as a MIMO model
⎡ ⎣YY12 Y3 ⎤ ⎦ = ⎡ ⎣ HH1121 HH1222 HH1323 H31 H32 H33 ⎤ ⎦ ⎡ ⎣XX12 X3 ⎤ ⎦ + ⎡ ⎣NN12 N3 ⎤ ⎦ (1)
where [Y1Y2Y3]T and [X1X2X3]T represent received signal and original signal, respectively.
[N1N2N3]T is the noise of the system. The coefficients in the matrixH is the channel response
for each path of the transmission. In order to obtain the channel coefficients, the time-multiplexing training sequences (TS) are added in front of the transmit data that can be expressed as
T1= ⎡ ⎣T S01 0 ⎤ ⎦ , T2= ⎡ ⎣T S02 0 ⎤ ⎦ , andT3= ⎡ ⎣ 00 T S3 ⎤ ⎦ . (2)
After the free space transmission, the originally zero parts in the training sequence become not zeros anymore; that is, the interferences from other channels and system noise will be included in the training sequence of the received signal. By dividing the received training data with the original ones, the channel response of each path can be calculated. Then, the channel matrix can be
Fig. 2. Experiment setup of 3×3 MIMO-OFDM VLC. estimated: H= ⎡ ⎣ YY1121//T ST S11 YY1222//T ST S22 YY2313//T ST S33 Y31/T S1 Y32/T S2 Y33/T S3 ⎤ ⎦ (3)
whereYi j denotes the received training sequence ofjth symbol of Rxi. Once the channel matrix is
obtained, the original data [X1X2X3]T can be recovered by multiplied the inverse matrixH-1 to the
left of the received signal to eliminate the interference from other channels, that is
⎡ ⎣XX12 X3 ⎤ ⎦ = ⎡ ⎣ HH1121 HH1222 HH1323 H31 H32 H33 ⎤ ⎦ −1⎡ ⎣YY12 Y3 ⎤ ⎦ . (4)
In this work, the MIMO training sequence is based on time-multiplexing in order to simplify the calculation of channel matrix H. As illustrated in the Tx side of Fig. 1, each training sequence of VLC signal consists of three durations, and only one duration carries the training data. For example, in the first time duration of training sequence at the Tx side, only LED1sends TS1, and the signals
sent from LED2and LED3at that time duration are zeros. Therefore, at the Rx side, Y11will be the
correct information; and Y21and Y31will be the interference from LED1to PD2and PD3. Hence, the
channel response from LED1to each PD can be simplified to the received training data divided by
TS1as shown in the first column of channel matrixHpresented in (3). Similarly, the values of the
other two columns ofHcan be calculated in the same way. As a result, the calculation of channel matrixHcan be simplified by using the time-multiplexing scheme (i.e., one channel sending training data, and the other channels sending zeros) when compared with the other scheme.
3. Experimental Demonstration
Fig. 2 presents the experiment setup of the 3 x 3 MIMO VLC system. First, three independent OFDM signals are created using offline MATLAB program with the parameters: fast Fourier transform (FFT) size of 256, cyclic prefix (CP) of 3.03%, and the subcarrier number of 30. Here, the bit-loading tech-nique is used in order to enhance the spectral efficiency; and the modulation order (using different quadrature-amplitude-modulation, QAM) is from 2 to 7 depending on the signal-to-noise ratio (SNR) of each subcarrier. In the experiment, the 3 independent electrical OFDM signals are generated by two arbitrary waveform generators (Tektronix, AWG7122B and AWG7082C) with sampling rate of 625 MSample/s. Here, two arbitrary waveform generators are used since each arbitrary waveform generator only has two ports and these two arbitrary waveform generators are synchronized. Then, they are applied to our developed front-end circuits, which contain pre-equalizers, electrical ampli-fiers and bias-tees for directly modulating the white LEDs with the distance of 12 cm between each LED as shown in Fig. 2. Three lenses are used to focus the light. The LED is from CreeR (XR-E).
OPA2677) and a bias-tee. The circuit layout and the detail parameters of the pre-equalizer circuit, such as the resistance and capacitance values can be found in [16]. In the experiment, each LED requires a DC bias voltage to turn-on, and the electrical OFDM data is applied together with the DC bias via a bias-tee circuit built inside the Tx. The DC bias voltage applied to the Tx is constant. As the OFDM is modulated at MHz range, during the experiment, we do not notice any flicking in illumination. Therefore, the modulation used here is the DC biased optical OFDM (DCO-OFDM).
In the beginning, the modulation levels of all OFDM subcarriers are set to 16-QAM, and then after the VLC MIMO experiment, the SNR of each subcarrier is measured. According to the measured SNR values, the modulation level of each subcarrier can be updated based on the bit-loading algorithm described in [9], so that the data rate of the system can be increased while satisfying the forward error correction (FEC) BER requirement. In the bit-loading, the subcarrier with higher SNR can carry higher modulation level (such as 64-QAM or 128-QAM). Finally, the new OFDM signal after bit-loading is updated in the arbitrary waveform generator, which is used to drive the LED system again.
After 1 m free space transmission, the three independent VLC signals are received with three PDs respectively. Instead of using three lens in front of each PD (known as non-imaging MIMO), we demonstrate imaging MIMO that only one lens is needed to focus three lights. Since the lens at Rx is used to focus the lights, the PDs can be closer to each other and thus the size of Rx will be more compact than non-imaging MIMO system. A blue filter can be included to attenuate the optical signal to avoid the PD saturation and to extend the analog bandwidth. Here, the Rx consists of PIN PDs (Hamamatsu, S10784) and trans-impedance amplifiers (TIA, Texas Instruments, OPA657). The PIN PD has the photosensitive area size of 7 mm2, optical-to-electrical response bandwidth
of 250 MHz, responsivity of 0.45 A/W, and dark current of 0.01 nA @ reverse bias voltage of 2.5 V. It has the spectral response range of 340 to 1040 nm. Afterwards, a real-time oscilloscope (Tektronix, CSA7404) with sampling rate of 1.25 GSample/s is utilized to receive the electrical signals from PDs for offline signal processing. First, the 3×3 MIMO demultiplexing as described in Section II is used to eliminate the interference of the signals from other channels. Then the recovered signals are demodulated and analyzed respectively. SNR and bit error rate (BER) measurement of each channel are performed in this paper. During the experiment, we do notice no flickering in illumination.
4. Results and Discussion
Before setting up the MIMO VLC system, the modulation bandwidth of Tx has been measured. Generally, the modulation bandwidth of LED is quite narrow and the whole system data rate will be limited. By using an electrical front-end Tx circuit which consists of a pre-equalizer integrated with an amplifier and a bias-tee, the modulation bandwidth of the system can be broadened widely. The detail implementation of the Tx circuit can be found in [14]. By adjusting the resistance and capacitance in order to match with the impedance of the LED, the available modulation bandwidth of the Tx can be extended as shown in Fig. 3. In this measurement, an electrical signal generator drives the VLC system with frequencies sweeping from 0 to 120 MHz, and the measured response by spectrum analyzer is illustrated in Fig. 3.
In the experiment, the OFDM signal consists of 30 OFDM subcarriers, and each is modulated in QAM format. Fig. 4(a) compares the measured SNRs of different OFDM-QAM subcarriers, and it also compares the SNRs of the received signals after MIMO demultiplexing. It can be observed that SNR drops dramatically at high frequencies. The low SNR is mainly caused by the limited modulation bandwidth of our VLC system and deficient optical power after 1 m transmission distance. Here, the SNR of Rx3 received signal is lower than other channels. The reason may be the imperfect match of impedance of LED so that the performance of the received signal will be a little worse. The corresponding modulation order of each subcarrier is shown in Fig. 4(b). The data rates of each channel are 258.79, 393.07, and 351.56 Mbit/s, respectively. The total data rate of
Fig. 3. Measured Tx modulation bandwidth without and with the front-end circuit.
Fig. 4. (a) Measured SNR of each VLC channel after MIMO demultiplexing and demodulated. (b) Corresponding bit-loading can be achieved in the bandwidth.
Fig. 5. Experimental constellation diagrams of received signal at Rx2, showing the constellation dia-grams of the subcarriers with modulation order of 3 (8-QAM) to 7 (128-QAM).
SNR. Finally, the average BER of the three VLC MIMO channels can be calculated. The BERs of each Rx are 2.4 × 10−3, 3.6 × 10−3 and 2.5 ×10−3, respectively. All of them satisfy the 7%
forward error correction (FEC) threshold of BER = 3.8 ×10−3. In addition, Fig. 5(a)–(e) shows the experimental constellation diagrams of received signal at Rx2, showing the constellation diagrams of the subcarriers with modulation order of 3 (8-QAM) to 7 (128-QAM).
5. Conclusions
VLC can be a potential candidate for the next generation wireless communications, and simultane-ously providing indoor lighting. The phosphor white-light LED has the advantages of low-cost, long lifetime and conserving energy. However, the narrow modulation bandwidth of phosphor LED is a big issue for future high speed communication. In order to enhance the data rate of the VLC system, in this work, we demonstrated a 3×3 imaging MIMO VLC system. Besides, pre-equalizer circuit was used to extend the bandwidth of the LED Tx from 14 MHz to 103 MHz. OFDM with bit-loading algorithm was used to further improve the spectral efficiency. By using the proposed scheme, the original about 1 MHz bandwidth commercially available phosphor white light LED can achieve 1 Gbit/s data rate over 1-m free-space transmission distance, satisfying the 7% FEC threshold of BER = 3.8 × 10−3. The SNR and bit-load of the OFDM signal were also studied.
References
[1] C. W. Chow, C. H. Yeh, Y. Liu, and Y. F. Liu, “Digital signal processing for light emitting diode based visible light communication,”IEEE Photon. Soc. Newslett., vol. 26, no. 5, pp. 9–13, Oct. 2012.
[2] Y. C. Chi, D. H. Hsieh, C. T. Tsai, H. Y. Chen, H. C. Kuo, and G. R. Lin, “450-nm GaN laser diode enables high-speed visible light communication with 9-Gbps QAM-OFDM,”Opt. Exp., vol. 23, pp. 13051–13059, 2015.
[3] W. Y. Linet al., “10m/500Mbps WDM visible light communication systems,”Opt. Exp., vol. 20, pp. 9919–9924, 2012. [4] C. H. Changet al., “A 100-Gb/s multiple-input multiple-output visible laser light communication system,”J. Lightw.
Technol., vol. 32, pp. 4723–4729, 2014.
[5] I. C. Lu, C. H. Yeh, D. Z. Hsu, and C. W. Chow, “Utilization of 1-GHz VCSEL for 11.1-Gbps OFDM VLC wireless communication,”IEEE Photon. J., vol. 8, no. 3, Jun. 2016, Art. no. 7904106.
[6] C. W. Hsuet al., “Visible light positioning and lighting based on identity positioning and RF carrier allocation technique using a solar cell receiver,”IEEE Photon. J., vol. 8, no. 4, Aug. 2016, Art. no. 7905507.
[7] Y. Liuet al., “Light encryption scheme using light-emitting diode and camera image sensor,”IEEE Photon. J., vol. 8, no. 1, Feb. 2016, Art. no. 7801107.
[8] H. L. Minhet al., “100-Mb/s NRZ visible light communications using a postequalized white LED,”IEEE Photon. Technol. Lett., vol. 21, no. 15, pp. 1063–1065, Aug. 2009.
[9] J. Vuˇci´c, C. Kottke, S. Nerreter, K. D. Langer, and J. W. Walewski, “513 Mbit/s visible light communications link based on DMT-modulation of a white LED,”J. Lightw. Technol., vol. 28, pp. 3512–3518, 2010.
[10] T. Fath and H. Haas, “Performance comparison of MIMO techniques for optical wireless communications in indoor environments,”IEEE Trans. Commun., vol. 61, no. 2, pp. 733–742, Feb. 2013.
[11] C. W. Hsu, C. W. Chow, Y. I. Lu, Y. L. Liu, and C. H. Yeh, “Demonstration of high speed imaging 3×3 MIMO-OFDM visible light communication system,” inProc. IEEE Photon. Conf., USA, 2016, Paper WP27.
[12] A. Burton, M. Hoa Le, Z. Ghassemlooy, E. Bentley, and C. Botella, “Experimental demonstration of 50-Mb/s visible light communications using 4 by 4 MIMO,”IEEE Photon. Technol. Lett., vol. 26, no. 9, pp. 945–948, May 2014. [13] A. H. Azhar, T. A. Tran, and D. O’Brien, “A Gigabit/s indoor wireless transmission using MIMO-OFDM visible-light
communications,”IEEE Photon. Technol. Lett., vol. 25, no. 2, pp. 171–174, Jan. 2013.
[14] S. Rajbhandariet al., “Imaging-MIMO visible light communication system usingμLEDs and integrated receiver,” in
Proc. Globecom Workshop, 2014, pp. 536–540.
[15] K. Cai and M. Jiang, “Multi-user MIMO-OOFDM imaging VLC system with PD selection,” inProc. IEEE 83rd Veh. Technol. Conf., pp. 1–5, May 2016, DOI: 10.1109/VTCSpring.2016.7504360.
[16] C. H. Yeh, C. W. Chow, H. Y. Chen, Y. L. Liu, and D. Z. Hsu, “Investigation of phosphor-LED lamp for real-time half-duplex wireless VLC system,”J. Opt., vol. 18, 2016, Art. no. 065701.
[17] R. A. Shafik, M. S. Rahman, and A. R. Islam, “On the extended relationships among EVM, BER and SNR as performance metrics,” inProc. Int. Conf. Electr. Comput. Eng., 2006, pp. 408–411.
