A frequency-shift-keying (FSK) visible light communication (VLC) system is proposed and demonstrated using advertisement light-panel as transmitter and mobile-phone image sensor as receiver. The developed application program (APP) in mobile-phone can retrieve the rolling shutter effect (RSE) pattern produced by the FSK VLC signal effectively. Here, we also define noise-ratio value (NRV) to evaluate the contrast of different advertisements displayed on the light-panel. Both mobile-phones under test can achieve success rate > 96% even when the transmission distance is up to 200 cm and the NRVs are low.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Due to the shortage of radio-frequency (RF) spectrum to meet the rapidly increasing bandwidth demands for wireless and mobile communications, using visible spectrum has raised considerable interest . Visible light communication (VLC) is evolving from pure research topics [2–7] to practical applications [8–10]. Light emitting diode (LED) is favorable transmitter (Tx) for VLC. Positive-intrinsic-negative (PIN) photo-diode (PD) and avalanche photo-diode (APD) are generally used as the VLC receivers (Rx). As personal devices such as mobile-phones and laptop computers are very common, and they are embedded with complementary-metal-oxide-semiconductor (CMOS) based cameras. It would be interesting using these CMOS image sensor for detecting VLC signal. However, due to the limited frame rates; using these cameras will significantly limited the VLC data rate. Many efforts have been made to solve this issue, including using tailor-made CMOS sensor with embedded PD ; however, it may not be low cost or commercially available in the market shortly. Another possible solution is to receive the VLC signal based on the rolling-shutter effect (RSE) in the CMOS based image sensor [12–15]. During the RSE operation, the light detection of pixel row is enabled sequentially. If the light source is modulated at the speed faster than the frame rate but slower than the pixel row activation time, bright and dark stripes are recorded by the camera.
In this work, we report a practical implementation of a VLC system using advertisement light-panel and CMOS image sensor based camera. The Tx is a commercially available light-panel used for advertising; and the Rx is a mobile-phone CMOS camera. The electrical modulation signal used to drive the light-panel is generated by using an Ardunio microprocessor, which is low-cost and practical. Here we also propose and demonstrate using frequency-shift-keying (FSK) modulation with parity check for error correction in this VLC system. Previous work of VLC system based on light-panel and mobile-phone CMOS camera used 4-pulse position modulation (4PPM) ; however, the experimental error was as high as 39.68%. By using the proposed FSK VLC signal encoding with decoding algorithms, higher tolerance can be observed in mid-to-low-end mobile-phones, which may have unstable frame rates. Longer transmission distance can be achieved when compared with previous work of light-panel and CMOS camera based VLC using on-off-keying (OOK) modulation . Detail designs of the proposed encoding and decoding algorithms for the FSK VLC signal with parity check are presented. We also define a noise-ratio value (NRV) to evaluate the contrast of different advertisements on the light-panel. Packet success rate measurements are performed at different NRVs and transmission distances. Both mobile-phones under test can achieve success rate > 96% even the transmission distance is up to 200 cm when the NRVs are low.
2. Experiment and algorithm
Figure 1(a) shows the proposed architecture of the VLC system using advertisement light-panel and CMOS image sensor. In the experiment, the FSK signal is generated by using an Ardunio microprocessor connected to the light-panel. 12-V direct-current (DC) bias is needed for the light-panel. The light-panel is commercially available with the size of 577 mm x 878 mm. A PMMA diffuser is located at the center of the light-panel and an 22W LED light bar (Epistar) is used to provide backlighting. In practical implementation, the Ardunio microprocessor can be located at the head office (HO) to drive several light-panels in a building simultaneously.
Figure 1(b) shows the structure of our proposed FSK VLC packet. It consists of three units and an “End” symbol. In each unit, there is a preamble, two data symbols and a parity. The preamble is used for the packet synchronization. Ten FSK frequencies are employed in our design to represent different symbols in the FSK VLC packet. As shown in Table 1, the FSK frequency used in the preamble is 781 Hz, which is the highest frequency. All the FSK modulation frequencies is much higher than 100 Hz ; hence the light flickering will not detected by human. The preamble is followed by data symbol 1 and data symbol 2. Each data symbol consists of 1-bit block code and two 2-bit data, as shown in Table 1. The last part of the VLC packet is the parity, which is obtained by performing XNOR operation of the data symbol 1 and data symbol 2. For example, if logic bit of (1000)2 is transmitted, it will first be divided into two groups, so that data symbol 1 is (10)2 and data symbol 2 is (00)2. For the parity, (10)2 XNOR (00)2 = (01)2 operation will be performed. Hence, the FSK frequencies in the VLC packet is 781 Hz (preamble), 600 Hz (data symbol 1), 520 Hz (data symbol 2), 650 Hz (parity). Since the mobile-phone CMOS image sensor is not able to detect the actual frequency of the FSK signal; the preamble frequency is used as the reference as shown in Table 1; and other relative FSK frequencies with respected to the preamble frequency can be obtained. In our design, due to different mobile-phone cameras will have different frame rates and shutter speeds, we consider the relative stripe width in the RSE pattern, which is the reciprocal of the relative frequencies. As a result, the relative stripe widths in the above VLC packet example is: 1 (preamble), 1.30 (data symbol 1), 1.50 (data symbol 2), 1.2 (parity). Here, each symbol takes 60 ms; and the symbol rate is 16.67 symbols per second; and the effective bit rate is 15.38 bit/s. The relatively low data rate can be enough to transmit promotion leaflet in brick-and-mortar stores, products' information in exhibitions, or positioning information in supermarkets. By using FSK VLC signal, much higher tolerance can be observed in mid-to-low-end mobile-phones, which have unstable frame rates. Longer transmission distance can be achieved as the illuminance level of the light-panel can be low. As shown in Table 1, ten-frequency options are used to represent different symbols. Higher data rate is possible by using more frequency options; so that more data symbols can be obtained. It is also possible to reduce the symbol time to increase the data rate.
Figure 2 shows the implementation layout of the proposed FSK VLC decoding scheme. The VLC packet is captured by the CMOS image sensor via our developed APP. The CMOS image sensor in the mobile-phone will not capture the whole image screen at the same time, but the pixel rows in the image sensor are enabled sequentially, named as RSE . When the light-panel Tx is modulated on and off, bright and dark stripes representing the FSK signals are captured in the image frame due to the RSE. Insets 2(a) and 2(b) show the image frames captured from the FSK signals of 781 Hz (the preamble) and 411 Hz (End symbol) respectively. We can clearly observe that the stripe widths are different in the two image frames. After this, 30 image frames are combined for signal decoding. The raw image files are converted into grayscale, where 255 means full brightness and 0 means darkness. A column matrix (1080 x 1) is selected from each frame for signal demodulation, and the selection conditions are illustrated in . Then, adaptive thresholding  is used for the stripe width (number of pixel occupied) calculation. From the obtained strip width, the actual frequency of the FSK signals can be retrieved from Table 1. After the FSK frequencies are obtained, the preamble as well as the VLC packet payload can be identified. Finally we measure the VLC packet success rate after different VLC wireless transmission distances. The defined NRV is used to estimate the contrast of advertisements. As shown in Eq. (1), where Ii,j and Ii,j’ are the grayscale value of a pixel of the picture on light-panel without and with the advertisement. i, j and rows, cols are vertical co-ordinate, horizontal co-ordinates, maximum numbers of the row and column respectively. In this experiment, NRVs of 0% (no advertisement), 39.56% and 70.21% are evaluated at the wireless transmission distances of 100 cm, 150 cm and 200 cm.
3. Results and discussions
As mentioned before, different mobile-phone cameras will have different frame rates and shutter speeds, we consider the relative stripe width in the RSE pattern in our proposed decoding algorithm. In the experiment, two mobile-phones: mobile-phone 1 (ASUS ZenFone 3) and mobile-phone 2 (HTC One A9s) are used. Both are using Android 6.0. Table 2 shows the measured strip widths of the two mobile-phones at different FSK modulation frequencies. At the preamble frequency of 781 Hz, the stripe widths of mobile-phone 1 and mobile-phone 2 are 58.7 pixels and 57 pixels respectively. Although each mobile-phone will have different stripe widths due to different shutter speeds (i.e. higher shutter speed will result in fewer pixels per stripe), by calculating the relative stripe widths at different FSK frequencies with respected to the preamble frequency, the obtained relative stripe widths are nearly the same.
Figures 3(a)-3(c) show an example of the 30 image frames (taken by the APP using mobile-phone 1 in 1 s at a distance of 100 cm) and the combined FSK VLC packet at NRV = 0%, 39.56% and 70.21% respectively. The red curves in the combined FSK VLC packet is the extreme value averaging (EVA) thresholding used . Although the three cases are measured at the same wireless transmission distance of 1 m, the grayscale values at NRV of 0% and 39.56% are around 150. When the advertisement with NRV of 70.21% is used, the grayscale value drops to around 100. We can also observe from the 30 image frames shown in Fig. 3(c) that the contrast is low. This will increase the difficulty to decoding the FSK signal and hence decrease the packet success rate.
Figures 4(a) and 4(b) show the measured VLC packet success rate at different NRVs and at different wireless transmission distances. We can observe that both mobile-phones can achieve success rate > 96% even the transmission distance is up to 200 cm when the NRVs are low. Figure 5 shows the corresponding illuminance at different transmission distances and at different NRVs. The illuminance at 200 cm when NRV of 39.56% (FSK modulation is on) is only 250 lux. When the NRV is 70.21%, the low contrast of the advertisement significantly increase the difficulty to demodulate the FSK packet; hence reducing the success rate when the transmission distance is > 100 cm. This can also be observed in both mobile-phones. As high NRV advertisement (dark picture) will affect the received signal-to-noise ratio (SNR), and affect the received VLC signal performance. This will limit the transmission distance. This is a limitation of the proposed VLC scheme. Hence, the NRV of the advertisement should be estimated before the VLC application. If the NRV is too high, some part of the picture should be reserved without advertisement in order to provide effective VLC.
We proposed and demonstrated a FSK VLC system using advertisement light-panel and CMOS camera. We also discussed the encoding and decoding algorithms for the FSK VLC signal. Here, we also defined NRV to evaluate the contrast of different advertisements displayed on the light-panel. Both mobile-phones under test can achieve success rate > 96% even when the transmission distance is up to 200 cm and the NRVs are low.
Ministry of Science and Technology, Taiwan, (MOST-106-2221-E-009-105-MY3) Aim for the Top University Plan, and Ministry of Education, Taiwan.
References and links
1. S. Wu, H. Wang, and C. H. Youn, “Visible light communications for 5G wireless networking systems: from fixed to mobile communications,” IEEE Netw. 28(6), 41–45 (2014). [CrossRef]
2. H. H. Lu, Y. P. Lin, P. Y. Wu, C. Y. Chen, M. C. Chen, and T. W. Jhang, “A multiple-input-multiple-output visible light communication system based on VCSELs and spatial light modulators,” Opt. Express 22(3), 3468–3474 (2014). [CrossRef] [PubMed]
3. areB. Janjua, H. M. Oubei, J. R. D. Retamal, T. K. Ng, C.-T. Tsai, H.-Y. Wang, Y.-C. Chi, H.-C. Kuo, G.-R. Lin, J.-H. He, and B. S. Ooi, “Going beyond 4 Gbps data rate by employing RGB laser diodes for visible light communication,” Opt. Express 23(14), 18746–18753 (2015). [CrossRef] [PubMed]
4. C. H. Chang, C. Y. Li, H. H. Lu, C. Y. Lin, J. H. Chen, Z. W. Wan, and C. J. Cheng, “A 100-Gb/s multiple-input multiple-output visible laser light communication system,” J. Lightwave Technol. 32(24), 4723–4729 (2014). [CrossRef]
5. Z. Wang, C. Yu, W. D. Zhong, J. Chen, and W. Chen, “Performance of a novel LED lamp arrangement to reduce SNR fluctuation for multi-user visible light communication systems,” Opt. Express 20(4), 4564–4573 (2012). [CrossRef] [PubMed]
6. T. C. Wu, Y. C. Chi, H. Y. Wang, C. T. Tsai, Y. F. Huang, and G. R. Lin, “Tricolor R/G/B laser diode based eye-safe white lighting communication beyond 8 Gbit/s,” Sci. Rep. 7(1), 11 (2017). [CrossRef] [PubMed]
7. H. Haas, “Visible light communication,” Proc. OFC2015, pp. 1–72.
8. C. W. Chow, R. J. Shiu, Y. C. Liu, Y. Liu, and C. H. Yeh, “Non-flickering 100 m RGB visible light communication transmission based on a CMOS image sensor,” Opt. Express 26(6), 7079–7084 (2018). [CrossRef] [PubMed]
9. C. H. Yeh, C. W. Chow, H. Y. Chen, J. Chen, and Y. L. Liu, “Adaptive 84.44-190 Mbit/s phosphor-LED wireless communication utilizing no blue filter at practical transmission distance,” Opt. Express 22(8), 9783–9788 (2014). [CrossRef] [PubMed]
11. I. Takai, S. Ito, K. Yasutomi, K. Kagawa, M. Andoh, and S. Kawahito, “LED and CMOS image sensor based optical wireless communication system for automotive applications,” IEEE Photonics J. 5(5), 6801418 (2013). [CrossRef]
12. C. Danakis, M. Afgani, G. Povey, I. Underwood, and H. Haas, “Using a CMOS camera sensor for visible light communication,” Proc. OWC’12, 1244–1248. [CrossRef]
13. H. M. Tsai, H. M. Lin, and H. Y. Lee, “Rollinglight-universal camera communications for single LED,” Proc. Annu. Int. Conf. Mobile Comp. and Netw. 317–320 (2014).
15. C. W. Chen, C. W. Chow, Y. Liu, and C. H. Yeh, “Efficient demodulation scheme for rolling-shutter-patterning of CMOS image sensor based visible light communications,” Opt. Express 25(20), 24362–24367 (2017). [CrossRef] [PubMed]
16. H. Aoyama and M. Oshima, “Visible light communication using a conventional image sensor,” Proc. CCNC 103–108 (2015). [CrossRef]
17. C. W. Chow, R. J. Shiu, Y. C. Liu, C. H. Yeh, X. L. Liao, K. H. Lin, Y. C. Wang, and Y. Y. Chen, “Secure mobile-phone based visible light communications with different noise-ratio light-panel,” IEEE Photonics J. 10(2), 7902806 (2018). [CrossRef]
18. P. Luo, M. Zhang, Z. Ghassemlooy, H. L. Minh, H. M. Tsai, X. Tang, L. C. Png, and D. Han, “Experimental demonstration of RGB LED-based optical camera communications,” IEEE Photonics J. 7(5), 7904242 (2015). [CrossRef]