High-speed quasi-balanced detection OFDM in visible light communication

Yuanquan Wang; Nan Chi; Yiguang Wang; Rongling Li; Xingxing Huang; Chao Yang; Ziran Zhang

doi:10.1364/OE.21.027558

1. Introduction

Visible light communication (VLC) based on light emitting diodes (LEDs) has drawn significant research interest because it can simultaneously support illumination and communication. It has been widely investigated in [1–5]. As far as we known, the highest data rates of uni-directional and bi-directional transmission are 3.4-Gb/s with a distance of 30cm [1] and 875-Mb/s with a distance of 66cm [2], respectively. However, one of the technical challenges is the inherent narrow bandwidth existed in both types of LEDs: red–green–blue (RGB) LED and phosphor-based LED. To overcome the limitation of bandwidth, most of the researchers consider orthogonal frequency division multiplexing (OFDM) technique as one of the promising modulation schemes for its high spectral efficiency (SE), and it has been widely employed in VLC.

But the second-order nonlinearity distortion in OFDM introduced by direct-detection (DD) is another drawback. Several approaches have been reported to reduce the distortion [6–9]. In [6], Lowery has proposed an offset single sideband (SSB) OFDM scheme to allocate sufficient guard band such that the signals and the intermodulation distortion (IMD) are no overlapping. In [7], a baseband SSB-OFDM scheme is proposed. However, SSB-OFDM cannot be realized in VLC system. Moreover, the SE in [6] and receiver sensitivity in [7] are sacrificed. In [8], iterated distortion reduction is proposed. It has better SE, but with a burden of computational complexity. In [9], asymmetrically-clipped optical (ACO) OFDM is proposed. In this scheme only odd subcarriers are filled so that the IMD will be located at even subcarriers. But it will lose efficacy in the case of frequency deviation.

Balanced detection (BD) is an efficient solution to this problem, but the LED is an incoherent source that BD cannot be employed in optical domain. In this paper, we proposed a novel quasi-balanced detection (QBD) technique in OFDM VLC system. By employing opposite signals to odd and even consecutive symbols, balanced detection with one single detector can be realized in electrical domain. This scheme has both advantages of balanced detection and direct detection, and can also be used in other similar cases such as in intensity modulation (IM) /DD OFDM short-range, low-cost optical fiber system. Using this scheme, the second-order intermodulation distortion and direct current (DC) can be eliminated, and the sensitivity of receiver can be improved. We achieved physical data rate of 2.1-Gb/s at a distance of 2.5 m. The BER performance can be largely enhanced by 22.2dB and 20.8dB compared with DDO-OFDM and ACO-OFDM, respectively. Therefore, QBD-OFDM outperforms the latter two schemes. In the remainder of this paper, we will give the detailed principle of QBD-OFDM. We will also investigate the feasibility in RGB-LED VLC system.

2. Principle of QBD-OFDM

The QBD-OFDM is mainly realized in electrical domain. The block diagram of this scheme is depicted in Fig. 1. The OFDM signals are divided into several blocks, and in each block, there are two symbols. The signals in the $2^{n d}$ symbol are opposite with the $1^{s t}$ one in the same block.

Fig. 1 Conceptual diagram of QBD-OFDM.

Download Full Size | PDF

In this scheme, the baseband OFDM signals in the $k^{t h}$ block are given by

E_{2 k - 1} (t) = \sum_{m = 1}^{N} c_{m} e^{j 2 π f_{m} t}

E_{2 k} (t) = - \sum_{m = 1}^{N} c_{m} e^{j 2 π f_{m} t}

where

N

is the number of subcarriers,

c_{m}

is the information symbol at the

m^{t h}

subcarrier, and

f_{m}

is the frequency of the subcarrier. And the OFDM signals which modulated on the LED light through bias-tee can be expressed as

s_{2 k - 1} (t) = (V_{0} - V_{a}) e^{j 2 π f_{0} t} + α e^{j 2 π (f_{0} + f_{1}) t} \cdot E_{2 k - 1} (t)

Where

V_{0}

and

V_{a}

are respectively the bias voltage and reversal voltage of LED. The coefficient

α

is used to describe the ratio of OFDM band strength related to the main carrier.

f_{1}

is up-converted frequency of the OFDM signals. After transmitting through the indoor environment, the OFDM signals considered frequency deviation and phase noise can be approximated as

r_{2 k - 1} (t) = (V_{0} - V_{a}) e^{j (2 π (f_{0} + Δ f) t + ϕ (t))} + α e^{j (2 π (f_{0} + f_{1} + Δ f) t + ϕ (t))} \cdot E_{2 k - 1} (t) + n_{0, 2 k - 1} (t)

Of which

Δ f

is the deviation frequency,

ϕ (t)

is the phase noise, and the

n_{0} (t)

is the Additive White Gaussian Noise (AWGN). After detected by the square-law detector, the photo currents for the

k^{t h}

block can be approximated written as

I_{2 k - 1} (t) = {| r_{2 k - 1} (t) |}^{2} + n_{r, 2 k - 1} (t) = I_{s, 2 k - 1} + I_{b, 2 k - 1} + I_{n l, 2 k - 1} + I_{n, 2 k - 1}

\begin{array}{l} I_{n, 2 k - 1} = 2 α Re {e^{- j (2 π (f_{0} + Δ f) t + ϕ (t))} n_{0, 2 k - 1} (t) \sum_{m_{1} = 1}^{N} c_{i} e^{- j 2 π f_{m} t}} + {| n_{0, 2 k - 1} (t) |}^{2} \\ + n_{r, 2 k - 1} (t) + 2 α (V_{0} - V_{a}) Re {e^{- j (2 π f_{0} t + ϕ (t))} n_{0, 2 k - 1} (t)} \end{array}

I_{s, 2 k - 1} = 2 α (V_{0} - V_{a}) Re {e^{j 2 π Δ f t} \sum_{m_{1} = 1}^{N} c_{i} e^{j 2 π f_{m} t}}

I_{b, 2 k - 1} = {| V_{0} - V_{a} |}^{2}

I_{n l, 2 k - 1} = {| α |}^{2} \sum_{m_{1} = 1}^{N} \sum_{m_{2} = 1}^{N} m_{2}^{*} m_{1} e^{(j 2 π (f_{m_{1}} - f_{m_{2}}) t)}

I_{2 k} (t) = {| r_{2 k} (t) |}^{2} + n_{r, 2 k} (t) = - I_{s, 2 k - 1} + I_{b, 2 k - 1} + I_{n l, 2 k - 1} + I_{n, 2 k}

where

n_{r} (t)

is the noise introduced by the receiver. In Eq. (5), the first term is proportional to the original OFDM signal, the second term is DC, the third term is the signal-to-signal beating noise (SSBN), and the fourth term is the other noises listed in the right of Eq. (6). In Eq. (6), the first term is the signal-to-AWGN beating noise, the second term is the AWGN-to-AWGN noise, and the fourth term is proportional to the AWGN noise. The photo currents of

2^{n d}

in the

k^{t h}

block can be expressed in the same way in Eq. (10). By subtracted Eq. (5) and Eq. (10), we can obtain the photo currents of the

k^{t h}

block.

I_{k} (t) = I_{2 k - 1} (t) - I_{2 k} (t) = 2 I_{s, 2 k - 1} (t) + {I_{n, 2 k} - I_{n, 2 k - 1}}

Where the first term is the signal and the second term is the noise. From Eq. (11), we can find the second-order intermodulation distortion, DC can be totally eliminated, and the sensitivity of receiver can be improved by 3dB, thus the signal-to-noise ratio can be improved.

3. Experimental results and discussion

The block diagram of QBD-OFDM VLC system is shown in Fig. 2. In this demonstration, a commercial available RGB LED (Cree, red: 620nm; green: 520nm; blue: 470nm) generating a luminous flux of about 6lm used as the transmitters (TX) and an avalanche photodiode (Hamamatsu APD, 0.42 A/W sensitivity at 620 nm and gain = 1) used as the receiver (RX) are adopted. The OFDM signals which consists of 64 subcarriers are generated by arbitrary waveform generator (Tektronix, AWG710) under almost the same processor expect the quasi-balanced (QB) transforming at TX and QBD at RX with our previous work in [3]. We also divide each frequency band into several sub-channels, and the bandwidths of all sub-channels are 50MHz (starting from 6.25MHz). Up-sampling by a factor 10 is employed to smooth the signals, and the sample rate of AWG is 500MS/s. Subsequently, the multiplexed QAM-OFDM signals came from AWG are filtered by a low pass filter (LPF) and amplified by electrical amplifier (EA) ((Minicircuits, 25-dB gain). The electrical QAM-OFDM signals and DC-bias voltage are combined via bias tee, and applied to different LED chips. Through free-space transmission, lens (100-mm focus length) and R/G/B filter, the data is recorded by a commercial high-speed digital oscilloscope (OSC) (Agilent MSO9404A) with 500MS/s sampling rate. Due to the limited experiment conditions, two OFDM signals used as the sources of LED chips come from the same output of AWG.

Fig. 2 Block diagrams of proposed QBD-OFDM VLC system (CP: cylix prefix; P/S: parallel to serial).

Download Full Size | PDF

In order to render the LED work at the optimal condition, we should find the proper bias voltage. In this demonstration, the distance between TX and RX is 0.5m, and the electrical input power is fixed at 12dBm, which is relatively small to avoid reaching the saturation area of LED. There are two sub-channels (sub1:6.25~56.25MHz; sub2:56.25~106.25MHz) on each LED chip, and the modulation format are MQAM (red: 256QAM, 128QAM; green: 128QAM, 64QAM; blue: 128QAM, 128QAM) -OFDM according to the knowledge of indoor channel. The experimental results are depicted in Fig. 3, and the constellations are shown for OFDM signals of red chip. We can find the optimal bias voltages for red, green and blue chip are 2.0V, 3.9V and 3.7V, respectively.

Fig. 3 Measured BERs versus bias voltage.

Download Full Size | PDF

In Fig. 4 we report the BER performances versus transmission distances in DDO-OFDM and QBD-OFDM VLC system. In a real scenario, the common distance between TX and RX is below 3 meters. So in this demonstration, the transmission distances are varied from 0.5 m to 2.5 m with step of 0.5 m. The bias voltages of RGB-LED are fixed at their optimum bias points as measured above. Other experiment parameters are the same with the bias voltage measured experiment mentioned above. The overall physical data rate is 2.1-Gb/s (red:750-Mb/s; green:650-Mb/s; blue: 700-Mb/s), including the training sequences, redundancy of cyclic prefix, FEC overhead, and repetition of opposite symbols in each block. As shown in Fig. 4, we can find the BER performances are largely enhanced for all sub-channels by employing QBD-OFDM. At the distance of 0.5 m, the BERs of sub1, sub2 in red, green and blue chips are improved by 25.6dB, 31dB, 30.3dB, 25.8dB, 21.8dB and 19.3dB, respectively. The constellations of blue chip adopting QBD-OFDM and DDO-OFDM are inserted in Fig. 4, and it shows that the constellations of QBD-OFDM are much better than the DDO-OFDM’s.

Fig. 4 Measured BERs versus distance.

Download Full Size | PDF

At last, we make a comparison of DDO-OFDM, ACO-OFDM, QBD-OFDM and QBD-ACO-OFDM. For simplicity, we just utilize green LED chip, and only one sub-channel is adopted to avoid the cross-talk from other sub-channels. The modulation format is 256QAM-OFDM, and the distance between the transmitter and the receiver is varied from 0.5 m to 2.5m. The electrical spectra of these four different types of OFDM signals are depicted in Fig. 5. As shown in this figure, only the QBD-OFDM and QBD-ACO-OFDM can eliminate the intermodulation distortion. The effectiveness of ACO-OFDM is not significant. Compared with the unoccupied spectra of ACO-OFDM or DDO-OFDM ranging from 60MHz to 100MHz, we can find that the background noises of QBD-OFDM and QBD-ACO-OFDM are also reduced. The experiment results of BER performances versus transmission distance are shown in Fig. 6. It can be seen the QBD-OFDM and QBD-ACO-OFDM can achieve error-free performance with 7% FEC threshold after 2.5 m indoor environment delivery. At the distance of 0.5 m, the BER performance of QBD-OFDM can be enhanced by 22.2dB and 20.8dB, compared with DDO-OFDM and ACO-OFDM, respectively. If we combine ACO- and QBD-OFDM together (i.e. QBD-ACO-OFDM), the BER performance can be enhanced by 27.0dB and 25.6dB, compared with DDO-OFDM and ACO-OFDM, respectively. The received OFDM signal constellations inserted in Fig. 6 are shown for the four types of OFDM, respectively. It can be easily seen that the constellations of QBD-OFDM and QBD-ACO-OFDM can be clearly recognized. Noting that, pre-equalizations have been employed in all of the above experiments, and in order to obtain the optimal performance of overall system, power allocation between sub1 and sub2 is also adopted.

Fig. 5 Spectra of: (a) DDO- (b) ACO- (c) QBD- (d) QBD-ACO-OFDM.

Download Full Size | PDF

Fig. 6 BER performance for different types of OFDM.

Download Full Size | PDF

4. Conclusion

In this paper, we have reported a novel QBD-OFDM scheme in VLC system based on a commercially available RGB-LED and a single detector. With this scheme, the second-order intermodulation distortion can be eliminated efficiently. And the sensitivity of receiver can be improved by 3dB. We have achieved physical data rate of 2.1-Gb/s at a distance of 2.5m in the condition of the overall brightness below 20lux enabled by higher-order modulation, pre- and post-equalization. The effective data rate is reduced by half due to the opposite symbols in the same block. But the measured BER performance can be enhanced by 22.2dB and 20.8dB compared with conventional DDO-OFDM and ACO-OFDM, and BERs are always below the FEC limit of 3.8x10⁻³. Therefore, QBD-OFDM could be an effective technique to achieve high data rate in short range and low cost access network.

Acknowledgments

This work was partially supported by the STCSM (No.12dz1143000), and NHTRDP (973Program) of China (Grant No. 2010CB328300), NNSF of China (No. 61177071, No. 61250018), NHTRDP (863 Program) of China (2011AA010302, 2012 AA011302).

References and links

1. G. Cossu, A. M. Khalid, P. Choudhury, R. Corsini, and E. Ciaramella, “3.4 Gbit/s visible optical wireless transmission based on RGB LED,” Opt. Express 20(26), B501–B506 (2012). [CrossRef] [PubMed]

2. Y. Wang, Y. Shao, H. Shang, X. Lu, Y. Wang, J. Yu, and N. Chi, “875-Mb/s Asynchronous Bi-directional 64QAM-OFDM SCM-WDM Transmission over RGB-LED-based Visible Light Communication System,” in Opt. Fiber Commun. Conf. (OFC), Anaheim, CA, 2013, OTh1G.3. [CrossRef]

3. Y. Wang, Y. Wang, N. Chi, J. Yu, and H. Shang, “Demonstration of 575-Mb/s downlink and 225-Mb/s uplink bi-directional SCM-WDM visible light communication using RGB LED and phosphor-based LED,” Opt. Express 21(1), 1203–1208 (2013). [CrossRef] [PubMed]

4. Y. Wang, M. Zhang, Y. Wang, Y. Fang, L. Tao, and N. Chi, “Experimental demonstration of visible light communication based on sub-carrier multiplexing of multiple-input-single-output OFDM,” in Opto-Electronics and Communications Conference (OECC), Busan, 745–746 (2012). [CrossRef]

5. N. Chi, Y. Wang, Y. Wang, R. Li, and H. Shang, “Application of digital signal processing in high-speed visible-light communication system,” in SPIE OPTO, San Francisco, CA, 86460L, 2013.

6. A. J. Lowery, L. Du, and J. Armstrong, “Orthogonal frequency division multiplexing for adaptive dispersion compensation in long haul WDM systems,” in Opt. Fiber Commun. Conf. (OFC), Anaheim, CA, 2006, PDP39. [CrossRef]

7. I. B. Djordjevic and B. Vasic, “Orthogonal frequency division multiplexing for high-speed optical transmission,” Opt. Express 14(9), 3767–3775 (2006). [CrossRef] [PubMed]

8. W. Peng, X. Wu, V. R. Arbab, B. Shamee, J. Yang, L. C. Christen, K. Feng, A. E. Willner, and S. Chi, “Experimental demonstration of 340 km SSMF transmission using a virtual single sideband OFDM signal that employs carrier suppressed and iterative detection techniques,” in Opt. Fiber Commun. Conf. (OFC), San Diego, 2008, OMU1. [CrossRef]

9. W. Peng, X. Wu, V. R. Arbab, B. Shamee, L. C. Christen, J. Yang, K. Feng, A. E. Willner, and S. Chi, “Experimental demonstration of a coherently modulated and directly detected optical OFDM system using an RF-tone insertion,” in Opt. Fiber Commun. Conf. (OFC), San Diego, 2008, OMU2. [CrossRef]

High-speed quasi-balanced detection OFDM in visible light communication

Abstract

1. Introduction

2. Principle of QBD-OFDM

3. Experimental results and discussion

4. Conclusion

Acknowledgments

References and links

Cited By

Figures (6)

Equations (11)

Optics Express