We demonstrate a non-flickering 100 m long-distance RGB visible light communication (VLC) transmission based on a complementary-metal-oxide-semiconductor (CMOS) camera. Experimental bit-error rate (BER) measurements under different camera ISO values and different transmission distances are evaluated. Here, we also experimentally reveal that the rolling shutter effect- (RSE) based VLC system cannot work at long distance transmission, and the under-sampled modulation- (USM) based VLC system is a good choice.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Visible light communication (VLC) is a popular research topic recently [1–7]. Light emitting diodes (LEDs) have now been widely deployed in our daily lives, such as general lighting, advertisement light panel, traffic light, etc. These LEDs can be deployed as the transmitter (Tx) for VLC. Many literatures of the VLC systems use positive-intrinsic-negative (PIN) photodiode (PD) or avalanche photodiode (APD) for VLC receiver (Rx) [4–6]. As mobile-phone and car with embedded complementary metal oxide semiconductor (CMOS) image sensors are common nowadays, using them as VLC Rxs can reduce the deployment cost of VLC. One important merit of VLC is that it can provide lighting and communication simultaneously. Hence providing non-flicker Tx light sources are crucial. The modulation frequency of the LED should be higher than the detectable flickering frequency of human eyes, which is ~100 Hz . However the frame-rate of most CMOS image sensor is only 30 or 60 frames per second (fps). In order to receive these high data rate VLC signals by using the low frame rate CMOS image sensor, the rolling shutter effect (RSE) can be used . Recently, different techniques to enhance the performance of RSE based VLC have been reported [10–14]. During the operation of RSE, the pixel row of the CMOS image sensor is activated one by one; hence bright and dark fringes representing the LED “ON” and “OFF” can be captured in each image frame. In the RSE based VLC system, a large portion of the CMOS image sensor should be exposed by the LED light in order to provide sufficient number of bright and dark fringes for the data demodulation; hence the transmission distance is highly limited. Another non-flickering VLC system detected by CMOS image sensor is using under-sampled modulation (USM) schemes, such as under-sampled frequency shift on-off keying (UFSOOK) or under-sampled phase shift OOK (UPSOOK) [8, 15]. In the UPSOOK scheme, by employing a dual-LED with a designated mapping and framing method, the data rate of 150 bit/s for a range up to 12 m has been demonstrated . Besides, by using red-green-blue (RGB) LEDs with multiple-input–multiple-output (MIMO), a 60 m VLC transmission at data rate of 150 bit/s (3 x 50 bit/s) was achieved .
In this work, we demonstrate a recorded non-flickering 100 m long distance RGB VLC transmission based on CMOS image sensor. Detail description and explanation of the decoding algorithms for the RGB UFSOOK signal are presented. Here, we also experimentally study bit-error rate (BER) measurements under different camera ISO values and different transmission distances. By using the built-in filter of the CMOS image sensor (no additional optical color filter is needed), the data rate can be increased by using wavelength division multiplexing (WDM). We also experimentally reveal that the RSE based VLC system cannot work at long distance transmission, and USM based VLC system is a good choice.
2. Experiment and algorithm
Figure 1(a) shows the experimental setup of the non-flickering 100 m long distance RGB VLC transmission system. Three channels of data in the UFSOOK format generated using MATLAB program in a computer are applied to the R, G and B LEDs respectively via an arbitrary waveform generator (AWG, Tektronix AFG3252C). The AWG acts as a digital-to-analog converter (DAC) converting the logic signals into actual electrical waveforms. The R, G and B LEDs are mounted on the same circuit board as shown in the inset of Fig. 1(a), and each LED is direct-current (DC) biased at 5 V and electrically driven at 2.5 VPP UFSOOK signal. The peak wavelengths of the R, G and B LEDs are about 630 nm, 520 nm and 470 nm respectively. The typical spectral half-width of the R, G and B LEDs are 20 nm, 30 nm and 20 nm respectively. The RGB VLC signals are received by the CMOS image sensor (Canon EOS 650D). The 100 m free-space transmission distance is limited by our building; and it can be observed that much longer transmission distance is possible when higher camera ISO values are used. The CMOS camera has the frame rate of 50 fps, shutter speed of 1/3200 Hz and the video resolution of 1280 pixels x 720 pixels. The video-mode of camera is employed and the recoded length is 40; hence 2,000 frames are captured in each recorded video. The experiment was performed at in-door environment with background light of ~300 lux.
The implementation of the UFSOOK modulation format is shown in Fig. 1(b), in which the black curve represents the LED “ON” and “OFF”. Each R, G, and B LED transmits different packet-based VLC signals, and each packet consists of 1-symbol header and 499-symbol payload. Each symbol occupies two image frames during the detection. The header is used for synchronization and is modulated at a higher frequency. During the VLC packet decoding, the header will be first located, and from the view of the CMOS camera, the received intensity appears as “Half-ON” since the LED is modulated at a high frequency. This means if the camera captures two “Half-ON”, it is regarded as the header. As shown in Fig. 1(b), due to the different modulation frequency of logic 1 (purple dotted box) and logic 0 (gray dotted box), if the camera captures two identical sampled frames (both “ON” or “OFF”), it is regarded as logic 1. If the camera captures two different sampled frames (one “ON” and one “OFF”), it is regarded as logic 0. For the camera to capture two light “ON” frames or two light “OFF” frames to represent a logic 1, the LED modulation frequency should be an integer multiple of the camera frame rate (i.e. n x fps), where n is an integer. For the camera to capture one light “ON” frame and one light “OFF” frame, the LED modulation frequency should be (n ± 0.5) x fps. As the frame rate used is 50 fps; the selected LED modulation frequency for logic 1 and logic 0 are 200 Hz (4 x 50 fps) and 225 Hz (4.5 x 50 fps) respectively. Since 2 frames represent one bit, the data rate in each color channel is 25 bit/s.
Figure 2(a) shows the decoding algorithm of the RGB UFSOOK VLC signal. The video file consists of 2,000 image frames is first read-in. Then the co-ordinate of the LED light spot is located. This LED light spot can be tracked by using the region-grow mechanism . After this, the R, G and B color channels in each image frame are separated. After the co-ordinate of the light spot is located; the grayscale values of the neighbor pixels within +/−100 pixels in both vertical and horizontal directions (total 200 pixels x 200 pixels) will be added up in each image frame. Hence, a total grayscale value in each image frame is obtained. The header image frame will be located by searching for the “Half-ON” state. Then, the payload can also be captured. Thresholding will be applied to all the images frames in the video to identify the logic 1 and 0 for the BER calculation.
3. Results and discussions
Figure 2(b) shows the measured normalized intensity response of the CMOS camera under different LED modulation frequency exposures. It is observed that the intensity response is reduced by > 2 dB when the LED modulation frequency is > 1.8 kHz; which is selected as the header frequency to represent the “Half ON” state. As discussed in last section, the selected LED modulation frequency for logic 1 and logic 0 are 200 Hz and 225 Hz respectively. These two frequencies have similar intensity responses in the camera; and they are much faster than the detectable frequency (~100 Hz) of human eye. Hence the flicking light causing dizziness could be avoided.
Figures 3(a)-3(c) show the image frames illustrating light “OFF”, “ON” and “Half-ON” respectively. As shown in Figs. 3(b) and 3(c), the slightly deviated R, G and B light spots are observed at the center of the frame. This is due to the different locations of the R, G and B LEDs in the circuit board, as shown in the inset of Fig. 1(a). In this demonstration, only 3 RGB LEDs are used to simply the illustration; however, more LEDs can be used and the VLC transmission data rate can be increased by using spatial division multiplexing (SDM).
As mention in the above decoding algorithm, a total grayscale value (all the grayscale values of the 200 pixels x 200 pixels around the light spot) is added up in each image frame. Figure 3(d) shows the total grayscale values of different image frames (from frame 1 to frame 2000) of the R color channel; and Fig. 3(e) shows a magnified section from frame 844 to frame 856. We can observe that the header is at frame 851 and frame 852 with the grayscale values between 6000 and 7000; and this illustrates the “Half-ON” state. Frame 845 and frame 846 show the light “ON” and light “OFF” states, representing a logic 0. Frame 847 and frame 848 show both light “OFF” states, representing a logic 1. Figure 3(f) and 3(g) show the total grayscale values in from frame 1 to frame 2000 for the G and B color channels respectively. We use the same algorithm for the logic detection.
Figures 4(a)-4(c) show the measured BER performances of the R, G and B VLC transmissions respectively under different free-space distances and at different ISO values. In each figure, the BER performances of only one color LED is turned on, and then all the RGB color LEDs are turned on are included for comparison. ISO value of the CMOS camera is the level of sensitivity of image sensor to the available light. A higher ISO value setting increases the CMOS image sensor sensitivity. For the R color channel (RGB case) as shown in Fig. 4(a), the BER satisfies the forward-error-correction (FEC) requirement (BER = 3.8 x 10−3) at the transmission distances up to 100 m when the ISO = 1600. If the ISO of 200 is used, 55 m free-space transmission can be supported in the R color channel at the RGB case; while 60 m transmission can be supported for the R color channel at the R only case. The increase in transmission distance at the R only case is due to the reduction of interferences caused by the other color channels. Similar transmission performances can be observed in both G and B color channels as shown in Fig. 4(b) and 4(c) respectively.
As discussed before, in order to receive non-flickering high data rate VLC signals by using the low frame rate CMOS image sensor, two main schemes: (i) RSE and (ii) USM can be used. In this section, we experimentally reveal that the RSE based VLC is not a good choice at long transmission distance although the data rate of RSE based VLC can be much higher. The experimental setup is similar to that shown in Fig. 1(a); however the RGB LEDs point light source is replaced by a LED white-light panel (60 cm x 90 cm) for better illustrating how a complete VLC packet cannot be received properly. The LED white-light panel consists of an LED light bar (Epistar) at the side to provide backlighting with output power of 22 W; and a PMMA diffuser at the center. The same CMOS camera with the same setting in previous experiment is used. Figures 5(a) and 5(b) show the image frames captured by the CMOS image sensor at the transmission distance of 3 m and 7 m respectively. We can clearly observe that when the transmission distance increases, the occupied area of the light panel image captured in the CMOS image senor becomes smaller. Since the RSE is produced at the CMOS camera, the RSE pixel width is unchanged when the transmission distance increases. As a result, when the transmission distance increases, a complete VLC packet cannot be observed in an image frame. The BER performance analysis of the RSE based VLC system at different data rates and different transmission distances are reported in . Although the RSE can provide a higher data rate, the transmission distance is limited; and USM is a good choice for long distance VLC transmission.
Recently mobile-phone and car equipped with high-resolution CMOS image sensors are common. This motivates using CMOS cameras for VLC. We demonstrated a non-flickering 100 m long distance RGB VLC transmission based on CMOS image sensor. Detail description and explanation of the decoding algorithms for the RGB UFSOOK signal are presented. We also experimentally studied BER measurements under different camera ISO values and different transmission distances. We also experimentally revealed that the RSE based VLC system cannot work at long distance transmission, and USM based VLC system is a good choice.
Ministry of Science and Technology, Taiwan (MOST-106-2221-E-009-105-MY3); Aim for the Top University Plan; Ministry of Education, Taiwan.
References and links
1. 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]
2. 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]
3. H. H. Lu, C. Y. Li, H. W. Chen, Z. Y. Yang, X. Y. Lin, M. T. Cheng, C. K. Lu, and T. T. Shih, “45 Gb/s PAM4 transmission based on VCSEL with light injection and optoelectronic feedback techniques,” Opt. Lett. 41(21), 5023–5026 (2016). [CrossRef] [PubMed]
4. J. R. D. Retamal, H. M. Oubei, B. Janjua, Y. C. Chi, H. Y. Wang, C. T. Tsai, T. K. Ng, D. H. Hsieh, H. C. Kuo, M. S. Alouini, J. H. He, G. R. Lin, and B. S. Ooi, “4-Gbit/s visible light communication link based on 16-QAM OFDM transmission over remote phosphor-film converted white light by using blue laser diode,” Opt. Express 23(26), 33656–33666 (2015). [CrossRef] [PubMed]
5. 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]
6. C. H. Yeh, L. Y. Wei, and C. W. Chow, “Using a single VCSEL source employing OFDM downstream signal and remodulated OOK upstream signal for bi-directional visible light communications,” Sci. Rep. 7(1), 15846 (2017). [CrossRef] [PubMed]
7. 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]
8. 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, 7904242 (2015).
9. C. Danakis, M. Afgani, G. Povey, I. Underwood, and H. Haas, “Using a CMOS camera sensor for visible light communication,” in Proceedings of OWC’12 (IEEE, 2012), pp. 1244–1248. [CrossRef]
11. K. Liang, C. W. Chow, and Y. Liu, “Mobile-phone based visible light communication using region-grow light source tracking for unstable light source,” Opt. Express 24(15), 17505–17510 (2016). [CrossRef] [PubMed]
12. W. C. Wang, C. W. Chow, L. Y. Wei, Y. Liu, and C. H. Yeh, “Long distance non-line-of-sight (NLOS) visible light signal detection based on rolling-shutter-patterning of mobile-phone camera,” Opt. Express 25(9), 10103–10108 (2017). [CrossRef] [PubMed]
13. 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]
14. J. Shi, J. He, J. He, R. Deng, Y. Wei, F. Long, Y. Cheng, and L. Chen, “Multilevel modulation scheme using the overlapping of two light sources for visible light communication with mobile phone camera,” Opt. Express 25(14), 15905–15912 (2017). [CrossRef] [PubMed]
15. R. D. Robert, “Undersampled frequency shift on-off keying (UFSOOK) for camera communications (CamCom),” in Proceedings of WOCC’13 (IEEE, 2013), pp. 645–648. [CrossRef]
16. P. Luo, Z. Ghassemlooy, H. L. Minh, X. Tang, and H. M. Tsai, “Undersampled phase shift on-off keying for camera communication,” in Proceedings of WCSP’14, (IEEE, 2014). [CrossRef]
17. W. C. Wang, C. W. Chow, C. W. Chen, H. C. Hsieh, and Y. T. Chen, “Beacon jointed packet reconstruction scheme for mobile-phone based visible light communications using rolling shutter,” IEEE Photonics J. 9(6), 1 (2017). [CrossRef]