Road-to-Vehicle Communication Using LED Traffic Light

Transcription

1 Road-to-Vehicle Communication Using LED Traffic Light Mitsuhiro Wada, Tomohiro Yendo, Toshiaki Fujii and Masayuki Tanimoto Department of Electrical Engineering and Computer Science Graduate School of Engineering Nagoya University Furo-chn, Chikusa-ku, Nagoya, , Japan Abstract- We propose a parallel wireless optical communication system for road-to-vehicle communication, which uses a LED traffic light as a transmitter and a high-speed camera as a receiver. The newly proposed system which enables multichanneling in LED two dimension arrangement and spatially dividing them. LED transmitters arranged in the shape of a plane are modulated individually, and a camera is used as a receiver for demodulating the signals by using image processing techniques. We performed examination of the transmission speed according to the proposed system in this paper. The actual proof experiment of the proposed system was conducted under the laboratory conditions. We checked that it communicates with a speed of 2.78kbps while 192 LED are spatially divided into eight groups and the high-speed camera set to 5OOfps. 1. NTRODUCTON Light-emitting diodes (LEDs) are increasingly used in traffic lights, message boards, signal devices, or other means of signalizing on the road. They are superior to incandescent lights due to their low power consumption, long life expectancy, high visibility, high tolerance to humidity, and minimal heat generation. LEDs are semiconductor devices that are capable of fast switching ON and OFF. This means LEDs can be used for communication purposes. Therefore, in addition to the normal function of being an indication and illumination device, LED traffic light can be used as a transmitter. Research on wireless optical communication using LED traffic light is performed in [ 1][2]. LED can transmit location information, safe driving support information, and other information from the road. Since light goes straight on, high directional communication is possible. So, transmitting different information for every lane of a road is possible. As for wireless optical communications, it is common to use a photodiode as a receiver and to use LEDs as a receiver. Photodiodes can not separate and detect the light from two or more LED transmitter. Therefore, in a typical wireless optical communication method, all LEDs of the transmitter are modulated by the sane signal. As a result pardkl communication is impossible. Recently, a parallel wireless optical communication method which uses a camera as a receiver is proposed in [2][3]. n this method, LED transmitters arranged in the shape of a plane are modulated individually, and a camera is used as 3 receiver to demodulate the signals by using image processing techniques. Parallel communication Transmit Different Data A ---*,rn Fig..,YY Y Y Y, The overview of the proposed system Hi gh-speed Camera as a Receiver is possible with a camera as a receiver. n this method, information is transmitted by many parallel LEDs, so this method gives the possibility of high speed communication. n this paper, a parallel wireless optical communication system using LED traffic light and high-speed camera is proposed for road-to-vehicle communication. Frame rate of commonly used video cameras is 30 frame per second (fps), while the frame rate of high-speed camera is hundreds to tens of thousands fps. n this system, we use a high-speed camera as a receiver to improve the transmission speed. We performed examination of the transmission speed according to the proposed method in this paper. The demonstration experiment of the proposed method was conducted under the laboratory conditions. 11. SYSTEM DESCRPTON OF THE PROPOSED SYSTEM The proposed system consists of an intelligent LED traffic light and a high-speed camera based receiver. Location information, safe driving support information, and other information from road are transmitted by the intelligent LED traffic light. The vehicle passing through an intersection with the high-speed camera based receiver can obtain this information. The scenario is shown in Fig.1. A. Parallel wireless optical communication The concept of parallel wireless optical communication is shown in Fig.2. LED transmitters arranged in the shape of 601 -

2 nome: Data- A 0 0 Om- E Data B Data C h Transmit Data in Parallel Lens High-speed Camera as a Receiver Fig. 4. High-speed Camera. Fig. 2. The-concept of parallel wireless optical communication. A high-speed camera as a receiver improve the transmission speed. A high-speed camera is usually very costly and bulky. Thus, a specific camera is needed and a protocol should be devised in order to adjust appropriate balance between cost and performance. D. The Estimate Transmission Speed The transmission speed of this proposal system is based on the following three elements. (1 ) The Amplitude of light signal (A). (2) The Number of LED Partitions (P). (3) The Signal Frequency (F). F is defined to be updating and flashing rate of the LED Traffic Light. The transmission speed of this proposal system is defined: Fig. 3. LED Traflic Light a plane send different data in parallel. A high-speed camera receives lights from LEDs as an image and decode it by using image processing techniques. After the transmitted data is decoded correctly, parallel data is converted to serial data. B. LED Traffic Light The response speed of LED is high (dozens of nanoseconds). f LEDs send data by fast switching ON and OFF, the switching is invisibk to the human eye. Since light from LED goes straight on, high directional communication is possible. C. High-speed Camera LED should be switched ON and OFF at a fast rate that human eye cannot detect. The flashing of the LED is undetectabk when the frequency of switching ON and OFF is over 50 Hz. However when the frequency is less than 100 Hz, flickers of light are visible. Therefore the frequency must be over 100 Hz. The capturing rate of the camera as receiver should be more than 200 frames per second, but frame rate of a commonly used video camera is 3Ofps. Thus a commonly used video camera is not suitable. While frame rate of high-speed camera is hundreds to tens of thousands fps, a high-speed camera is used to capture the image of the intelligent LED traffic light. (log, A PF (1) When the unit offis set to Hz, the unit of this speed is set to bit per second (bps). The response time of the existing LED used by traffic light ranges from 5 MHz to 50 MHz, which limits the transmitting speed by one LED. n this system, while, data is transmitted in parallel by using many LEDs, transmitting data in the speed of Gbps is possible. But frame rate of a high-speed camera is not so high to receive such speed of data. For example, frame rate of the high-speed camera ranges from 500 fps to 30,000 fps. Presently, communicating in 2 Mbps is possible when A = 2, P = 200 and F = 10,000 Hz. Further improvement in the camera speed is expected. E. Character of Data Transmission Channel n this system, transmitted data is decoded from an image, so the luminance information on each LED within a image is needed. f each LED is captured vividly, the luminance information can be got correctly. When some LEDs overlap in defocused image, and when the area of LED Traffic Light is too small in captured image, decoding the transmitted data is difficult. This means data transmission channel is degradated. The following three parts are the degradation factor of data transmission channel quality: (a)attenuation of LED light. (b)reduction of the number of pixels to which LED is reflected. (c)leds overlap in defocused image. (b) and (c) *

3 h;l Light Source Camera Sensor P ane Lens Camera Sensor Plane Optical Axis ~ * 4 d d Center of LED Projection d Fig. 5. Ray tracing method by the optical system. are the degradation factor particular to this system which uses the camera as a receiver. We formulate these degradation by ray tracing method. ;;...il The following terms are defined: / Distance between axis of lens and light source; *!/ U The angle formed by ray and axis of lens; ,.._ r and 71 are the parameter of light source. Perpendicular distance from the camera lens to the object; p, p Pixel width of the camera sensor plane; y 1 P i n Number of pixels of the camera sensor plane;..-.-._....., r The effective radius of lens; f Focal length of the camera lens; d Distance between the lens and the camera sensor plane;.%~..#b Fig. 6. LED projection on the camera sensor plane. B Angle of view; h Distance from axis of lens on the camera sensor plane;..-..,(..u The angle formed by ray and axis of lens on the camera 8/2 e + sensor plane; w A diameter of a LED projection on the camera sensor,_.. f plane; ; Distance between two LEDs np : Size of Camera Sensor Plane k Distance between centers of two LED projections Fig. 7. Definition of angle of view. A data transmission channel is modeled as shown in Fig. 5. The transition ray is formed in (2). (; ;) And the refraction of ray is shown in (3). ( -l+ ;) By these matrices, the relations between (h, U) and (h, U ) are shown in (4). (. tayu ) = ( ) ( - + ) ( i, ) ( tanu (2), 4w1 p distance between centers of two LED projections k should be bigger than the diameter of the LED projection w to avoid the overlap problem. k is as follows. d k = -k (6) (3) 1 The condition that LEDs do not go over is the next expression. d d d -k 1 > 2r(l ) - J Especially, it shows the next expression when the lens is (41 ) focused on infinite distance (d f). Fig. 6 shows a LED projection on the camera sensor plane. k 2 2r (8) w is derived from d, 1, f and T. This expression means that distance between two LEDs should d d w = (hi be bigger than the effective aperture of camera lens. ~ = 2r(l ) (5) f l On the other hand, quantities of light are insufficient when Decoding is difficult when a LED projection is overlapped the effective aperture is small. So, optimum effective aperture with aother LED projection on the camera sensor plane, The is equal to the distance between nearest two LEDs. (7)

4 Fig. 8. Consecutive image 1 captured by high-speed camera Fig. 9. Consecutive image 2 captured by high-speed camera. Distance between centers of two LED projections should be bigger than the size of two pixel. Definition of angle of view 0 is as follows. (Fig. 7.) Expression (12) is led from expression (O) and expression (1 ). This is the conditional expression for decoding. Wighresolution or telescopic lens are necessary for correct decoding DETECTON AND TRACKNG THE PORTON OF LED TRAFFC LGHT N TKE MAGES To obtain the transmitted data, detecting and tracking the. portion of LED Traffic Light in the images are needed. Two consecutive images captured by the high-speed camera are similar except the LED traffic light. The differencing process is an effective method for detecting and tracking the LEDs within the images to separate LED Traffic Light from background. Fig. 8 and Fig. 9 are consecutive images captured by the highspeed camera and Fig. O is a differented image of Fig. 8 and Fig. 9. The movement of other vehicles in the scene and the camera own movement cause noise in Fig. 10. This noise is linear and shape of LED Traffic Light is not linear. By using this character, LED Traffic Light is detected and tracked. v. BASC EXPERMENT We made the basic experimental system and convert the LED Traffic Light for transmitter. Fig. O. A differenced image of Fig. 8 and Fig. 9. A. Experiment System Architecture Fig. 11 shows the experiment system. ) Transmitter: The mounted transmitter is composed of LED panel of LED Traffic Light, their drive circuits, and the FPGA. The LED panel is divided into eight partitions (see Fig. 12). One partition is assigned the task of the standard LED. The FPGA in the transmitter divides data into seven groups. 2) Receiver: The receiver composed of a ens and highspeed camera Ad a PC. A PC extracts and decodes the received data by image processing. B. Experiment Cunditiuns The actua1 proof experiment of the proposed system was conducted under the laboratory conditions. The transmitter and the receiver were fixed, Distance between the transmitter and the receiver was four meters. The ampitude of the light signal A was 3. The number of LED partitions P was seven. One partition is assigned to the standard LED used as the

5 < standard of synchronization. Signal frequency f is 250 Hz. n the proposed system, the frame rate of high-speed camera is twice the LED update rate, so the frame rate of the high-speed camera is 500 fps. C. Frequency of LEDs mid,high-speed Camercr Fig. 13 shows the problem that transmitted data captured as wrong signal. An image is the integration value of light. When a camera captures a part of growing LED and a part of not growing LED in a frame, the luminance of LED in captured image is not correct. Therefore, the problem shown in Fig. 13 occurs when the frame rate of high-speed camera and the LED update rates are the same and the synchronization of high-speed camera and LED is shifted. n the proposed system, the frame rate of the high-speed camera is twice the LED update rate, and some LEDs used as the standard of synchronization, called standard LEDs, are set. The standard LEDs invert the signal, for example from ON to OFF or from OFF to ON, at every update of LEDs. To obtain the data, selection of the images is needed. The method of selection of the images is using the luminance of standard LED in the four sequential frames of images checked and compared. Then, maximal luminance image and the minimal one arc selected. Fig. 14 shows the method of selection of the images. n selected images. the standard LEDs continues switching On and OFF while one image is captured. Therefore, the problem shown in Fig. 13 is avoided. D. Luminance Modulation An image is the integration valuc of light as mentioned above. We apply luminance modulation in proposed system by LED light pulse-width modulation. Fig. 15 shows the concept of luminance modulation. The luminance is judged by comparing with the luminance of standard LED. E. Result Data transmission and reception was successful. Transmission speed was calculated by A = 3, P = 7 and f = 250, (log, 3) x 7 x 250 = 2782 (12) Transmission speed was 2.78 kbps. t checked that parallel wireless optical communication using LED Traffic Light and a high-speed camera was possible. V. CONCLUSON n this paper, a parallel wireless optical communication system using LED traffic light and high-speed camera for road-to-vehicte communication was proposed. n this system, LED traffic light used as transmitter and a high-speed camera used as receiver. LED traffic light in which LEDs modulated individually transmits data in parallel, and a camera demodulates the received data by using image processing techniques. Using this system, LED traffic light can transmit location information, safe driving support information, and the other information from road. Since light goes straight, high directional communication is possible. So, transmitting differcnt information for every lane of a road is possible.... FPGA..... i.: *:.: *:,'nput Data i.:. ; Output...: *: *:.:.. - i PC. c_ c_ LEO Traffro Light High-speed Camera U -. mage Processing... :-~ : Fig. 12. Fig... Expenmenr System Architecture The LED panel is divided into eight partitions

6 Fig. 14. receiver. Transmitter LED ON 2 Standard ON OFF ON ON L Camera j i-j j Receiver t*- i i i Captured mage of LED Fig. 13. Captured mage of LED 1 Captured as Wrong Signal 1 ON The transmitted data captured as wrong signal l Se ect Max i ma t um i nance image and mi n i ma um i nance image The timing relationship berween'the standard LED and the camera We estimate transmission speed. The transmission speed of this proposal system is defined by the amplitude of light signal, the number of LED partitions and the signal frequency. The expression is as follows: (log, A)PF To achieve a high transmission speed, the number of LED. partitions is an important factor. Character of data transmission channel of this system is discussed. n this system, transmitted data is decoded from an image, so character of data transmission channel is specified by the camera and the lens. Attenuation of LED light, reduction of the number of pixels to which LED are reflected and LEDs overlap in defocused image are the degradation factor of data transmission channel quality. We formulate a LED width in the captured image. The LED width becomes the standard of evaluation of the degradation of data transmission channel. We made a demo system and a basic experiment was conducted indoors. This experiment demonstrate that a parallel wireless optical communication system using LED traffic light and high-speed camera is possible. ACKNOWLEDGMENT The authors thank Dr. T. Yamazato, Dr. H. Okada of EcoTopia Science nstitute, Nagoya University, and Mr: K. Masuda of graduate school of engineering, Nagoya University. REFERENCES [ 11 M.Akanegawa, YTanaka and MNakagawa. Basic Srudy on Trufic nformarion Swrem Using LED Trafic Lights. EEE Trans. TS. vol. 2, No. 4, pp , Dec., Hugh Sing Liu and Grantham Pang, Positioning Beacon Syarem Using Digid Camera and LEDs, EEE Transactions on Vehicular Technology, vol. 52, No. 2, March [3] S.Miyauchi, T.Komine. T.Ushiro, S.Yoshimura, S.Hmyama and M.Nakagawa, Parullel Wirekss Opricul Cnmmlrnicnrion uring High Speed CMOS mage Sensor, ECE Technical Report VoLlW, No.81, CS2Do4- WOCS RN , May Luminance of LED Lumi name ON. _ JN-L- OFF t Light Pulse-width Modu a t i on Fig. 15. Luminance modulation *