Vol. 51 No. 3 899 913 (Mar. 2010) 1 2 1 1 1 GPS GPS GPS GPS GPS GPS 80 m 80 m 2 3 GPS 0 GPS GPS GPS 5 CGI NTT KDDI 98% A Pedestrian Positioning System Using Road Traffic Signs and Landmarks Tomoyuki Kojima, 1 Kazuya Yamane, 2 Masao Yanagisawa, 1 Tatsuo Ohtsuki 1 and Nozomu Togawa 1 Mobile-GPS is generally used for pedestrian positioning on mobile devices such as mobile phones and PDAs. Positioning errors of mobile-gps might be caused by several factors, such as multipath, however, positioning errors of mobile-gps have been not investigated sufficiently. In this paper, we first investigate positioning errors of mobile-gps at Takadanobaba station and its environs which have both urban and residential areas. Our investigation results show that positioning errors of mobile-gps can cause approximately 80-meter error at the maximum. Secondly we propose a highly accurate pedestrian positioning method using road traffic signs and landmarks. Our proposed method does not require any infrastructure construction as we already have infrastructure of road traffic signs, landmarks and mobile-gps on mobile devices. Assuming that a user is positioned at the traffic sign, our proposed method determine the user position by finding out several nearby road traffic signs and sending their colors and shapes to a server. Our method start with locating approximately position of a user using mobile-gps. Next, it locates user position by selecting road traffic sings and landmarks. Our method is implemented with CGI and investigated using mobile phones of NTT Docomo and au by KDDI. By this investigation, the accuracy of this method was 98% and we succeeded to confirm effectiveness of the proposed method through this evaluation investigation. 1. 2007 4 1) GPS Global Positioning System GPS GPS GPS GPS GPS 1 GPS m m 2) GPS m m 1 Department of Computer Science and Engineering, Waseda University 2 Department of Computer Science, Waseda University 899 c 2010 Information Processing Society of Japan
900 GPS GPS DGPS Differential GPS RTK-GPS Real Time Kinematic GPS 2) DGPS RTK-GPS GPS GPS Wi-Fi 3) RFID 4) M-CubITS 5) Wi-Fi PSP PlayStation Portable 1 7) 100 m DB RFID RFID M-CubITS 0 1 M-CubITS M GPS GPS GPS A-GPS Assisted GPS GPS 8) A-GPS GPS 2006 5 GPS 115 m 9) GPS 8) GPS 2.2 76 m 76 m 2 3 GPS GPS 0 1 1 1 PSP 2004 12 SCE 6) 1 Fig. 1 A road traffic sign. 3 1 GPS GPS 4 GPS GPS DB GPS 1 1 Perl CGI NTT KDDI SoftBank GPS 98%
901 2. GPS GPS GPS GPS GPS 2.1 0 2007 10 GPS 1m 10) 2 1 2 2 10 m 5m 5m 2.2 GPS GPS GPS GPS 5 GPS GPS GPS 2006 5 GPS 9) KDDI W51S 200 m 8) GPS 2006 5 GPS GPS NTT N905i KDDI W47T 1 NTT 2 KDDI GPS 2009 1 NTT 81 KDDI 48 2 3 NTT GPS 4 5 KDDI GPS 26.96 m 76.14 m 5m 26.96 m 1 76.14 m 1 2 GPS 1 GPS 76.14 m GPS 2.3
902 Table 1 1 Infrastructure of existing positioning systems. DB GPS 11) GPS 1) 1) DGPS 2) RTK-GPS 2) M-CubITS 5) M-CubITS IC 4) IC bluetooth 12) bluetooth Wi-Fi 3) 2 NTT GPS Fig. 2 Investigated points for mobile-gps positioning error using an NTT DoCoMo mobile phone. 3 Fig. 3 NTT GPS Mobile-GPS positioning errors using an NTT DoCoMo mobile phone. Table 2 2 Positioning accuracy of existing positioning systems. GPS 11) 500 m m GPS 100 m m 1) 50 m 100 m 1) 100 m 100 m 10 km DGPS 2) 100 m m RTK-GPS 2) cm cm M-CubITS 5) m IC 4) 20 m bluetooth 12) 10 m Wi-Fi 3) 100 m 4 KDDI GPS Fig. 4 Investigated points for mobile-gps positioning error using a KDDI mobile phone. 5 KDDI GPS Fig. 5 Mobile-GPS positioning errors using a KDDI mbile phone. DB 1 DB 2 GPS GPS GPS D-GPS RTK-GPS GPS M-CubITS IC bluetooth Wi-Fi GPS 1 DB GPS GPS DB GPS DB 1 2
903 2009 DB 5m GPS 5m 5m GPS GPS 13) (1) (2) GPS GPS GPS 5m GPS DGPS GPS GPS GPS 3. GPS 2.2 GPS 2.3 GPS 0 GPS 9 14) GPS 2011 4 100% 1) GPS 2008 812,638 15) 500 m 40 79 m 137 73% 109/137 1 2.7 DB DB DB DB DB DB DB DB DB
904 80 m 40 m 40 m 3.1 GPS 4 (1) GPS GPS GPS (2) GPS (3) (4) 6 DB DB DB 6 Fig. 6 Overall processing flow.
905 DB DB DB DB 3 DB DB DB DB DB DB DB 7 DB google map 16) 137 8 GPS 3.2 GPS GPS GPS GPS 200 m 2.2 GPS 80 m GPS 9) 200 m 2 5 200 m 3.3 GPS 7 DB Fig. 7 Area where we have constructed our database. 2 9),17),18)
906 Fig. 9 9 Example of frame colors and shapes of road traffic signs. 8 Fig. 8 Processing flow from the viewpoint of a user. (1) 17) (2) 17) (3) 17) (4) (5) 30 (6) (4) (6) 1 17) 2,700 (1) (5) (6) 100 3.3.1 9 9 18) 3.3.2 3.3.3 7 (1) 20 7 (2) 7 19) 3 7
907 Table 3 3 The frame colors and shapes of road traffic signs. 9 7 3.3.2 (1) (7) (1) 6 (a) (2) GPS 6 (b) (3) (4) 6 (c) (4) 3.3.3 6 (d) (5) 1 (6) 1 6 (a) (7) 2 (4) 6 (e) 2 1 1 2 m 2 1 1 1 5 10 Fig. 10 Number of road traffic signs placed nearby a crossroad in urban areas. 3.3.3 2 1 1 m 10 10 1 8 2 8 1 2 2 8 10 1
908 2 10 1 7 10 8 8 1 2 8 8 1 1 40 m 8 40 m 8 60 m 1 40 m 60 m (1) (2) 2 5 (3) 3.1 DB DB 3.1 DB DB 11 11 137 25 112 40 m 60 m (2) 3.4 11 Fig. 11 Simulation area. 1 3 1 2 1 4 3 3.5
909 Table 4 4 Number of the points where a road traffic sign is specified correctly in each step. 1+2+3 1+2 1+3 137 13 0 137 3.6 1 3.1 DB GPS 0 GPS GPS 3.1 DB DB 4 4 1 2 3 GPS 4 GPS GPS 1 GPS 3 2 4. GPS 4.1 Perl CGI 1 google static map 16) google static map 12 TOP 4.2 3.1 DB DB 3.1 DB DB NTT 137 KDDI 48 13 14 NTT KDDI 13 14 240,000 m 2 1 137 NTT 4.3 GPS GPS 1 TSPS Traffic Sign based Positioning System http://www.togawa. cs.waseda.ac.jp/tsps/cgi-bin/start.cgi 2009 5 NTT KDDI SoftBank GPS 2009 5 3.1 7
910 Fig. 13 13 NTT Investigated area using an NTT DoCoMo mobile phone. 12 CGI Fig. 12 Implemented system using CGI.
911 5 Table 5 Investigation results. GPS 5.4% 115 122 NA 10 20 N905i 100.0% 0 137 1.28 30 W47T 95.8% 2 48 1.30 29 15 1 Fig. 15 A road traffic sign which could not be found by users (1). 98% 14 KDDI Fig. 14 Investigated area using a KDDI mobile phone. 5m 5m 4.4 1 NTT N905i 2 KDDI W47T GPS 2008 12 2009 1 5 GPS 5m GPS 2.2 GPS 5 GPS 5.4% 2 2 15 16 15 16 16 16 16 2 100% 5 1.3 7 1
912 GPS 0 CGI NTT KDDI 98% 15 16 DB 21650238 16 2 Fig. 16 Road traffic signs which could not be found by users (2). 1 1 1 2 1 GPS 1 20 2 40 GPS GPS 76 m 2 5. GPS GPS 1) (May 2004). 2) GPS Vol.82, No.12, pp.1207 1215 (1999). 3) koozy. http://www.koozyt.com/ 4) 2 ITS pp.488 493 (2003). 5) M-CubITS ITS2009-28, pp.115 119 (2009). 6) PlayStation.com. http://www.jp.playstation.com/psp/ 7) http://www.zenrin.co.jp/ 8) FOMA NTT DoCoMo Vol.13, No.4, pp.14 19 (2006). 9) ITS2007-29, pp.33 38 (2007). 10) GPS 12 12 5 (2007). 11) GPS 2003-ITS-13, pp.25 32 (2003). 12) Bluetooth Beacons 2001-MBL-18, pp.69 74 (2001). 13) GPS H CAS97-112, pp.9 16 (1998).
913 14) 2009 R&D (2008). 15) 20 http://www.keishicho.metro.tokyo.jp/ 16) Google. http://www.google.co.jp/ 17) ITS2008-14, pp.37 42 (2008). 18) DICOMO2008 pp.844 852 (2008). 19) KICTEC. http://www.kictec.co.jp/ ( 21 5 24 ) ( 21 12 17 ) 1985 2008 1986 2009 1959 1981 1983 1986 1987 1990 IEEE ACM OR 1940 1963 1965 1980 LSI 1969 1994 32 IEEE CAS Society Guillmin-Cauer Prize Award 1974 Meritorious Service Award 1995 Golden Jubilee Medal 2000 2000 IEEE 3rd Millennium Medal VLSI I Layout Design and Verification North-Holland IEEE 1970 1992 1994 1997 VLSI 1996 9 1997 21 IEEE