Proposal of MAC concept for VL-ISC (Visible Light Image Sensor Communication)

Transcription

1 Proposal of MAC concept for VL-ISC (Visible Light Image Sensor Communication) Nobuo IIZUKA CASIO COMPUTER CO., LTD. / VLCC 1

2 Contents Characteristics of ISC New MAC concept of ISC The example of CASIO s ISC Conclusion 2

3 Next... Characteristics of ISC New MAC concept of ISC The example implementation of CASIO Conclusion 3

4 Image Sensor Communication (ISC) Spatial separation capability Brightness-Distance invariant law Providing Data simultaneous with Spatial position (Assumption: MAC mode is broadcast / Multi-pt to pt communication ) Noise X1,Y1 Data B... Data A Data B lens X2,Y2 Data A... About VLC using Arrayed PD / Image sensor doc.: IEEE < July > 4

5 Many applications of ISC List No.41 Positional measurement at Factory and warehouse!!! List No. 47 ITS Navigation!!! List No. 46 Guiding in spectacle outdoor List No.43 Robotics (Indicate and command) Application summary list: doc.: vlc-application-definitions-and-summary.xls Vehicle o Vehicle Pre crash safety Using spatial property in communication 5 List No.40 Positional sensor net

6 Motivation of ISC based on Visible Light characteristics ISC utilizes Visible Light advantages that should be considered 1. The short wavelength provides high spatial resolution It is not only for High data rate or many channels 2. Image formation lens is usable It provides Imaging capability. ISC solves Visible Light environment issue that should support A lot of strong ambient light noise/ interference Sun light is severe DC noise in outdoor. Indoor fluorescent lamp has kHz noise. 6

7 Next... Characteristics of ISC New MAC concept of ISC The example implementation of CASIO Conclusion 7

8 In a reductive point of view The structure is arrayed photo diode. Photo diode = channel = Is it the merely arrayed channels? We think No It brings a qualitative change Does it need the consideration for standard? We think Yes we should consider it 8

9 A new MAC concept of ISC Conventional Wireless Communication Upper layer ISC Upper layer Received data Received Received data Received data Position data Position Position Scene image Regarding space as Shared media MAC layer Necessary multiple access control PHY MAC layer Unnecessary multiple access control Necessary Spatial Consideration Discovery / Tracking function required ISC PHY Regarding space as Separated media For the new effect / new issue, we should consider it 9

10 Discovery and Tracking required function of ISC Discovery It is Channel (s) selection from many channels in the array. Many Signals and Many Interference Tracking optional function based on discovery It is like a hand-over method of mobile phone system. t 10

11 Next... Characteristics of ISC New MAC concept of ISC The example implementation of CASIO Conclusion 11

12 A realization of ISC-MAC Policy : Simple architecture, available device Performance of discovery have to be Stabilized in any environment. Wide range of scalability in data rate ultra low data rate (100bps) to low data rate ( 5K bps ) Suitable method for discovery : Consider about sending data for ISC. Receiving devise use a expansion of sending out data. 12

13 Data format with ISC expansion Add specific PN (Pseudo Noise) code field in a data block. The PN-code is a key of discovery. Every natural brightness fluctuation and the PN-code do not have a correlation each other. Other part of the block and the PN-code also do not have a correlation each other. data block For discovery & synchronization PN code field ( Header ) other part of data ( Payload ) 7-13bit length will be enough. 13

14 ISC expansion : receive side process Scene Image Signal Position Example PNcode= t t1 t2 t3 t4 t5 Image FIFO One signal and many interference Discovery and synchronization at the same time The load of the correlative calculation is same in both 1 signal and 1,000 signals. Matched filter of header PN code For discovery When there is signal 14

15 ISC Prototype of CASIO doc.: IEEE < > Sensor module Various display Program Depends on Application Scene image USB Received data(s) and Position(s) PC Fast COMS Image Sensor + ISC processer (fps-windowing controllable, Almost normal structure) 15

16 Confirmation experiment about eliminating interference VLC data signal 300Hz mod. Freq. Display output of application 300Hz freq. Signal That has no PN-code Handy light (intense DC) Interferences 16 CASIO COMPUTER CO., LTD. There is not the influence of interfering.

17 A demo movie Discovery and Tracking using proposed data format CASIO 2009.Sep. 2minutes. 17

18 The list of ISC examples Image sensor use existing sensor Application conscious CASIO (VLCC) TOSHIBA (VLCC) NEC (VLCC) Performance conscious Make special hardware SONY (VLCC) KEIO Univ. (VLCC) Shizuoka Univ. NAIST.etc These examples have discovery function. 18

19 Next... Characteristics of ISC New MAC concept of ISC The example of implementation of CASIO Conclusion 19

20 VLC MAC/PHY logical diagram ISC includes a new concept of MAC, but it may be possible by expansion of conventional MAC. From the viewpoint of time-to-market, the standardization of ISC should begin examination by the PHY of the low data rate. VLC MAC ISC MAC (around 500 to around 100k bit/s) Low Data Rate PHY for ISC High Data Rate PHY for PD 20

21 Conclusion doc.: IEEE < > New MAC concept was proposed and explained that spatial consideration is mandatory. Especially discover is essential. An effective method for discovery was explained. As ISC point of view, PHY modulation frequency should be considered in law data rate at first. ISC is one of the fundamental technology, and not just implementation matter. Spatial consideration for ISC should be discussed in TG7. 21

22 Appendix 22

23 10 July 2009 doc.: IEEE < > Brightness Distance Invariable Characteristics "Brightness doesn't depend on the distance" is a fundamental law in optics. Energy of 1 pixel L(distance) = P/S L n Irradiation energy to lens = P Image formation area = S The energy for 1 pixel is the same. Energy of 1 pixel = (P/n 2 ) / (S/n 2 ) Irradiation energy to lens = P/n 2 = P/S Image formation area = S/n 2 23

24 Stable decoding by one pixel selection According to Brightness-Distance invariable low, One Pixel Selection provides easy gain control and wide range distance. Intensity (Log scale) Noise floor 1000 One pixel Maximum Excessive input energy (image of many pixels) Total intensity of light signal image (Inverse square low) Intensity of One pixel A stable domain to use for communication Distance (Linear) 24

25 A rough confirmation experiment 5mmφ LED Under 1pixel size d=2m d=8m d=14m Pixel Intensity Of image center t 1LED d ISC measuring tool Distance: 2 14 (x7) Intensity: 1->2/3 Not 1/49! (7x7=49) 25

26 Example of FOV difference Image Sensor +/- 20 degree Flat. No half-value concern Photo-diode with lens Relative Intensity 0-100% Half value at +/-30 degree 26

27 Two FOV of ISC - 1 doc.: IEEE < > ex1) Device FOV: different / Pixel FOV: same Pixel FOV: xx degree ex)xga( ) Device FOV:46 degree Same distance 同じ面積 / 明るさの光源 ex)qvga( ) Device FOV:46 degree Pixel FOV: xx degree 27

28 Two FOV of ISC 2 doc.: IEEE < > Device FOV: same / Pixel FOV: different Pixel FOV: xx degree XGA( ) Device FOV: 46 degree Same distance Pixel FOV: xx degree QVGA( ) ¼ 解像度 Device FOV: 46 degree n 倍のパワー 28

29 Weak indirect light and a lot of interference ISC can receive by the feeble signal of the reflection. (indirect light). "Pasting up of the information" is possible with spotlights such as Art Gallery / Museum / Store. Spot down light Display output of application Information balloon in bottom of display 29

30 10 July 2009 Image Sensor : Spatial Separation Capability doc.: IEEE < > Robustness to Ambient Light noise. No interference even when the Multi Point light sources are received. Using IS Noise Power = 1 Data signal Power = 1 20 degree Device FOV Image sensor Selected element of array (Pixel in image sensor) Signal : Noise = 1:0 The noise is excluded by the spatial separation. Device FOV doesn't affects SNR directly. (Not Device FOV but Element FOV of array decides SNR.) *1 An appropriate array element can be properly selected. *2 Array and optical system have a sufficient resolution capability. Ex) Same device FOV 20 degree Using PD Noise Power = 1 20 degree Device FOV PD Signal : Noise = 1:1 Everything is accumulated. Data signal Power = 1 SNR deteriorates because the noise ratio Increases when Device FOV is getting larger. 30