IWAIT January 6-8, Pukyong National University, Busan, Korea International Workshop on Advanced Image Technology

Transcription

1 ISBN IWAIT International Workshop on Advanced Image Technology January 6-8, 2016 Pukyong National University, Busan, Korea Organized by The Korean Institute of Broadcast and Media Engineers (KIBME), Korea Co-Organized by Institute of Electronics Information and Communication Engineers (IEICE), Japan Institute of Image Information and Television Engineers (ITE), Japan Japanese Society of Precision Engineering (JSPE-IAIP), Japan National Science Council (NSC), Taiwan IEEE Broadcast Technology Society, Seoul Chapter IEEE Consumer Electronics Society, Seoul Chapter Electronics and Telecommunications Research Institute (ETRI), Korea Supported by LG Electronics, JINMYUNG BROADCAST, MPEG Forum The Korean Federation of Science and Technology Societies (KOFST) Busan Tourism Organization (BTO), Any Future Technology (AFT)

2 TECHNICAL PROGRAM P.2C-5 View synthesis using homography transformation for 360 VR service Gun Bang, Gwhang soon Lee ETRI, Korea P.2C-6 Multiple Sensorial Media Application Format Sang-Kyun Kim, Jung Yup Oh Myongji University, Korea P.2C-7 Standardization of Media-centric Internet of Things Sang-Kyun Kim Myongji Unversity, Korea P.2C-8 Near-Duplicate Video Copy Detection with Multi-Modal Video Signature Matching Jun-Tae Lee, Kyung-Rae Kim, Won-Dong Jang, Chang-Su Kim Korea University, Korea P.2C-9 BLE Beacon-based Position Tracking System in Android Platform Dong-Kyu Choi, Joong-Hwa Jung, Sang-Il Choi, Seok-Joo Koh Kyungpook National University, Korea P.2C-10 Implementation of CoAP/6LoWPAN over BLE Network for IoT Services Cheol-Min Kim, Hyung-Woo Kang, Seok-Joo Koh Kyungpook National University, Korea P.2C-11 LABORATORY MEASUREMENT TO PROVIDE THRESHOLD OF VISIBILITY FOR TERRESTRIAL 4K-UHDTV BROADCASTING BASED ON HEVC OVER DVB-T2 IN AWGN CHANNEL Sungho Jeon, Junghyun Kim, Jaekwon Lee, Sanghun Kim, Jeong-Deok Kim, Young-Woo Suh, Zungkon Yim KBS, Korea P.2C-12 RGB-D Image Segmenation Based on Random Walk with Restart Se-Ho Lee, Won-Dong Jang, Chang-Su Kim Korea University, Korea P.2C-13 Getting Higher Performance Using a Two-Layer Extreme Learning Machine Junhyuk Hyun, Jeonghyun Baek, Jisu Kim, Peyman Hosseinzadeh Kassani, Euntai Kim Yonsei University, Korea P.2C-14 Implementation of Integrated multi-media measurement and monitoring system Young-Woo Suh, Junghyun Kim, Sungho Jeon, Zungkon Yim, Sung-Choon Park KBS, Korea P.2C-15 Relaxation Based Matching of SIFT Keypoints Clustered by Gradients and Location Sunmin Lee, Yong Cheol Kim University of Seoul, Korea P.2C-16 Toward 3D Object Tracking without Jittering Minseok Kang, Jungsik Park, Jong-Il Park Hanyang University, Korea P.2C-17 The optimal bilateral remote control system under disaster environment Sun Lim, Young Wook Kim KETI, Korea P.2C-18 Multiple Player Tracking in Indoor Sports with Multi-cameras Yookyung Kim, Gi-Mun Um, Kwang-Yong Kim, Kee-Seong Cho, Hyungmin Kim*, Jong-Il Park* ETRI, *Hanyang University, Korea P.2C-19 Human Gait Prediction method using Microsoft Kinect Junghwan Kim, Doyoung Kim, Inwoong Lee, Jongyoo Kim, Heeseok Oh, Sanghoon Lee Yonsei University, Korea P.2C-20 Copyright Protection and Compact- and Secure-Transmission of Diagnosed Fundus Images Wannida SAE-TANG King Mongkut s University of Technology, Thailand IWAIT

3 BLE Beacon-based Position Tracking System in Android Platform Dong-Kyu Choi, Joong-Hwa Jung, Sang-Il Choi, Seok-Joo Koh* School of Computer Science and Engineering Kyungpook National University Daegu, Korea Abstract To provide various convenient services using smartphone, the interests in Bluetooth or Bluetooth Low Energy (BLE) have been increased. One of the useful functions using Bluetooth Beacon is user position estimation with the distance measurement. However, it is difficult to measure the actual distance and position in a simple way, because Bluetooth Beacon is vulnerable to interference. To overcome this problem, we proposes two optimized distance measurement schemes: the average distance scheme and the prune off scheme. In each proposed scheme, we calculates the average distance using all or some measured distance between Bluetooth Beacon and user and uses the calculated average distance as a distance value. Then, based on three distance values, we can obtain the user position estimated by three candidate schemes. From the performance analysis, we compare three candidate schemes in terms of accuracy of position estimation. As a result, we can see that the two proposed schemes show better accuracy than general scheme and prune off scheme has the best performance. Keywords Bluetooth Beacon, Positioning Tracking System, Bluetooth 4.0, BLE, Android I. INTRODUCTION With the growth of smartphone services, the interests in services and technologies using Bluetooth or Bluetooth Low Energy (BLE) [1-2] have been increased. The Bluetooth Beacon [3] using Bluetooth 4.0 is a device to provide shortrange wireless communication less than 50 meters, and it can solve several limitations of short-range wireless communication protocol such as NFC [4], GPS [5], and Wi-Fi [6]. The Bluetooth Beacon can be used in various indoor or outdoor environment, and also it can overcome the radio shadow area problem of GPS. In addition, the Bluetooth Beacon supports BLE, and thus it can reduce the power consumption and solve the high power consumption problem of Wi-Fi. Many related products have been developed, such as ibeacon of Apple [7], Eddystone of Google [8], Paypal beacon of Paypal, and etc. However, most of existing works have focused on the development of product. There is no study on the services to provide convenience by utilizing the Bluetooth Beacon. One of the useful functions of Bluetooth Beacon is to measure distance between Bluetooth Beacon and mobile device. For the distance measurement service using Bluetooth Beacon, it transmits the Bluetooth signals periodically. After that, several smartphones can receive the Bluetooth signal at the same time. On the reception of Bluetooth signal, smartphone sends the information, which is the Unique User Identifier (UUID) of Bluetooth Beacons and the measured distance to a server. To calculate the location of a user, at least three Bluetooth Beacons are required to uses trilateration. Then, the server sends the coordinates image resources and positions to the smartphone. However, there are some difficulties to measure distance using Bluetooth Beacon due to measurement error. If the measurement error is continuously occurred, it is hard to get the exact position of the user. To overcome this problem, in this paper, we proposes two optimized distance measurement schemes: the average distance scheme and the prune off scheme. In the average distance scheme, the accurate data is calculated by average of all measured values in each Bluetooth Beacon. By using an average of the all measured values, it is possible to reduce the measurement error. On the other hand, in the prune off scheme, the accurate data is calculated by average of selected some values. In this scheme, we checks the distribution chart and remove the values of both end sides. Then, we calculate the average using remained values. By discarding some error values, it can improve the accuracy of the calculated value. This paper organized as follow. Section II briefly introduce the measuring method using Bluetooth Beacon. Section III reviews general positioning tracking system using Bluetooth Beacon. In Section IV, we presents two proposed schemes. Section V analyzes and compares three candidate schemes. Section VI concludes this paper. II. MEASURING METHOD USING BLUETOOTH BEACON A. Distance Measurement Using Bluetooth Beacon To estimate a distance between smartphone and Bluetooth Beacon, the signal level and reference level of beacon are

4 needed. In the case of signal level, Bluetooth Beacon provides a RSSI (Received Signal Strength Indication) [9] value as a signal level. For example, we assume that packets advertised by Bluetooth Beacon are arrives with the -65dBm signal level calibration, and the output of the reference level is -59dBm. Since the signal level of -65dBm is weaker than -59dBm, the Bluetooth Beacon is likely to be at a distance greater than 1 meter. Therefore, we can calculate the distance between smartphone and Bluetooth Beacon using two values, which is RSSI and reference level. Fig. 1 shows the distance estimation function used in Android Beacon Library [10]. III. GENERAL POSITIONING TRACKING SCHEME In the general positioning tracking scheme, a smartphone receives only one distance value between user and Bluetooth Beacon at a time. Although various measurement errors can be occurred, there is no special error correction procedures. Fig. 1. Distance estimation function used in Android beacon library B. Position Measurement Using Bluetooth Beacon To estimate a position of user, the most commonly used scheme is triangulation algorithm. In this algorithm, the position is calculated by three different point. First, it measure the distance between the user and three different points. Then, it draws circles with a radius, which is the distance between the user and itself. At this time, the intersection of three circles is the position of user. However, if Bluetooth Beacons are used as a distance measurement method, there is no intersection of three different circles due to measurement error. So, in this paper, additional calculations are added. First, we draws three straight line using each two intersections of two circles. Then, there is one intersection among three lines. In this paper, we assumed that this intersection is estimated position of user. Fig. 2 shows the modified position measurement algorithm to considering the measurement error of Bluetooth Beacon. Fig. 3. Example of measured data in general positioning tracking system Fig. 3 shows the example of measured data in general positioning tracking system. In this figure, we set the distance between Bluetooth Beacon and user to 8-meters and measured 50 times. According to the figure, it is shown that the measured distance is very irregular due to measurement error of Bluetooth Beacon. In addition, this incorrect distance value has a bad effect on the positioning tracking system. Fig. 4 shows the estimated user position with incorrect distance values. Fig. 4. Example of estimated user position with incorrect distance values Fig. 2. Modified position measurement algorithm In the figure, the estimated position of user (red star) is so far from actual position of user (yellow star). As mentioned above, it is because the position estimated by each distance between user and Bluetooth Beacon that contains the measurement error. To provide an effective positioning tracking service using Bluetooth Beacon, the error of measurement error should be solved.

5 IV. PROPOSED POSITIONING TRACKING SCHEMES To overcome the measurement error of Bluetooth Beacon, we focus on the distribution of the measured values. Fig. 5 shows the distribution of the measured values of Bluetooth Beacon. In the figure, we set the distance between Bluetooth Beacon and mobile device to 8-meters and measure repeatedly 2000 times. B. The Prune Off Scheme In the prune off scheme, we measure the distance between Bluetooth Beacon and user several times at a time and draw distribution chart. In this paper, we set the number of distance measurements to 20. Then, we select the most gathered interval and remove other samples, which is within other intervals. In this scheme, the distance between Bluetooth Beacon and user is average value of measured distance in selected intervals. Fig. 7 shows the flow diagram of prune off scheme. Fig. 5. Distribution of the measured values of Bluetooth Beacon According to the figure, it is shown that the frequency of a value similar with actual distance is higher than other incorrect value. In addition, the measurement error is greater, the probability of occurrence is decreased. Based on this analysis, we proposes two Bluetooth Beacon-based positioning tracking schemes: average distance scheme and prune off scheme. A. The Average Distance Scheme In the average distance scheme, we measure the distance between Bluetooth Beacon and user several times at a time and calculate an average distance. In this paper, we set the number of distance measurements to 30. Then we set the average distance as a distance between Bluetooth Beacon and user. Fig. 6 shows the example of estimated distance using average distance scheme. Fig. 7. Flow diagram of prune off scheme First, in this scheme, it extracts twenty-measured distance between Bluetooth Beacon and user at a time. Then, it displays all measured distance value in the distribution chart. In this case, we assume that the interval of distribution chart is 0.5- meters. When the distribution chart is completed, it counts the number of values in each interval and selects an interval that has the biggest value. Then, these distance values in selected interval are stored at selected distance. Additionally, distance values nearby selected interval are stored at selected distance. After that, it calculates the average value using selected distance values and defines the average value as a distance value. Fig. 6. Exmaple of estimated distance using average distance scheme In the figure, we set the distance between Bluetooth Beacon and user to 8-meters and measured 50 times using average distance scheme. As a result, we can see that each estimated distance using average distance scheme is similar to the actual distance. It is because the distribution of measured distance is similar to normal distribution and average calculating operation offsets by large and small measurement errors. However, it still has a tendency in which the measured distance is little shorter than the actual distance. V. PERFORMANCE ANALYSIS For the performance analysis, we made up the experimental model with smartphone and three Bluetooth Beacons. Each distance between Bluetooth Beacon and smartphone is set to 8- meters. In other words, the smartphone is located at (0, 0) and three Bluetooth Beacons are located at (8, 0), (0, 8), and (-8, 0) respectively in two-dimensional plane. In this experimental environment, we measure and estimate three distances between each Bluetooth Beacon and smartphone. Then, based on three distances, we estimate the coordinate of smartphone and verify the accuracy of three candidate schemes by comparing the actual and estimated coordinates.

6 Fig. 8. Measured and estimated distance of three candidate schemes Fig. 8 shows the measured and estimated distance of three candidate schemes. In this figure, we can see that the general scheme has a large variation because of measuring error due to interference. On the other hand, the variations of two proposed schemes are smaller than general scheme. This is because the occurred distance error processing by the averaging. In addition, the prune off scheme can remove some large distance error value using distribution chart. Then, it calculates an average distance using only reliable distance values. As a result, the prune off scheme shows best accuracy among three candidate schemes. VI. COCLUSION In this paper, we propose two optimized distance measurement schemes: the average distance scheme and the prune off scheme. In the average distance scheme, the accurate data is calculated by average of all measured values in each Bluetooth Beacon. On the other hand, in the prune off scheme, the accurate data is calculated by average of selected some values. In this scheme, we will check the distribution of the measured values and remove the values of both end sides. Then, we calculate the average using remained values. From the performance analysis, we can see that the proposed two optimized distance schemes show better accuracy than general positioning tracking scheme because of processing of measurement error. However, in case of the prune off scheme, it reduces the number of sample distance value compared with the average distance scheme by removing some large measurement error values using distribution chart. So, it can reduce the overhead of sample distance measurement. Therefore, we can see that the prune off scheme has best performance and accuracy among three candidate schemes. ACKNOWLEDGMENT This research was supported by the MSIP (Ministry of Science, ICT and Future Planning), Korea, under the SW- Centered University Program (IITP-2015-R ) supervised by the IITP (Institute for Information& communications Technology Promotion) and by the BK21 Plus project (SW Human Resource Development Program for Supporting Smart Life) funded by the Ministry of Education, School of Computer Science and Engineering, Kyungpook National University, Korea (21A ). Fig. 9. Estimated user position of three candidate schemes Fig. 9 shows the position of the user on the twodimensional coordinates. As mentioned above, the general scheme has biggest distribution and worst accuracy, because it does not consider the measurement error of Bluetooth Beacon. On the other hand, two proposed scheme shows similar distribution and accuracy because of the distance error processing by using calculated average value. Comparing the two proposed schemes, the prune off scheme has lower number of extracted sampling distance value because of removal of some large measurement error values using distribution chart. As a result, despite the lower number of sample distance value, the prune off scheme can have similar result with average distance scheme and it means that the prune off scheme has the best performance. REFERENCES [1] Jin-Shyan Lee, A comparative study of wireless protocols: bluetooth, uwb, zigbee and wi-fi, 33rd Annual Conference of the IEEE, November [2] Gomez, C.; Oller, J.; Paradells, J. Overview and Evaluation of Bluetooth Low Energy: An Emerging Low-Power Wireless Technology, Sensors 2012, 12, [3] JoaKim Lindh, Bluetooth low energe beacons, in Texas Instruments, January 2015 [4] Near field communication, [5] Simon D. P. Williams, et al., Error analysis of continuous GPS position time series, Journal of geophysical research Vol.109 Issue B3, March 2004 [6] Andrei Croitoru, Costin Raiciu and Dragos Niculescu, Wi-fi mobility without fast handover, IETF-RFC [7] Andy Cavallini, ibeacon Bible 2.0, [8] Google Developers, [9] M. Sugano, T. Kawazoe, Y. Ohta, and M. Murata, Indoor localization system using RSSI measurement of wireless sensor network based on zigbee standard, Wireless and Optical Communication MultiConference, July 2006 [10] Android-beacon library,