IPSJ-CVIM

Size: px

Start display at page:

Download "IPSJ-CVIM"

爱尹
5 years ago
Views:

1 modeling of large-scale environment Laser-based multiple mobile robots using Yongjin Jeong, y Yumi Iwashita, y y y y y 3 3 CPS-SLAM.3% RGB-D. )) 5) 3)4) 3 7) 6) 3 8) Parent robot -axis attitude sensor y and Tsutomu Hasegawa y Ryo Kurazume have been proposing a high precision laser-based 3D measurement system We by multiple mobile robots. This system is composed of three mobile robots Child robot Child robot of a parent robot and two child robots. The parent robot is equipped consisting a 3D laser scanner, attitude sensor and a total station, and the child robots with 3, CPS-V equipped with corner cubes. The parent robot moves and stops repeatedly, are measures the 3D shape using the equipped laser scanner at several posi- and Meanwhile, the child robots also move and stop repeatedly, and act as tions. landmarks for the positioning of the parent robot. This paper presents mobile improvements of the proposed system by replacing and installing several some devices to make the positioning accuracy higher. The experimental results the system achieves quite high accuracy of the.3 % of target's size. show omni-directional sensing robot equipped with four RGB-D cameras and its The (Cooperative Positioning measurement experiments are also introduced. y University, Graduate School of Information Science and Electrical Engineering Kyushu y Kyushu University, Faculty of Information Science and Electrical Engineering cfl Information Processing Society of Japan

2 ο 3; m :3 resolution :5 = Angular + ppm Distance ±mm (angle) Accuracy (Applied Geomechanics Inc.) MD-9-TS ±5 ffi Range 9) GPS System, CPS) (IMU) 6).3% ) ) 3) ) 4) 5) Mapping SLAM Simultaneous Localization and Kinect Microsoft RGB-D Child robot Total station Parent robot Remote contoller LMS 5 (SICK AG) angle 7 ffi View :5 ffi /:5 ffi Auto-leveling system Laser range finder 6 CPS-VI 3 Child robot. resolution Angular error Systematic ±3mm :4 ffi Resolution : ffi (= 36 ) Repeatability Statistical error ±mm. 6 (TOP- CON, GPT-95A, ) AS- (SICK, ), ARS-36-HP Applied LMS-5, Inc., MD-9-TS 3 Geomechanics (TOPCON A3 ) TOPCON, RC-3) GPT-95A (TOPCON) frequency Scanning range Max. 5Hz/5Hz 5m Hysteresis : ffi (= ) 4 CPS) 5 9) 6) (Cooperative Positioning System, () () (3) Range Accuracy (distance) (4) 6) CPS 3 m 33.9m.97m.3% 6) IMU 3 cfl Information Processing Society of Japan

Turn table -axis laser range finder 6) CPS 6 CPS 7 CPS Slit-like laser 4 Measured local shape Measured local shape

Positioning of child robots Aligned model Child robots () Robot and move. () Robot measures the position of robot.

ICP(Iterative Point) 3 Closest r ψ φ (c) Positioning of parent robot (d) Measurement by LRF 6 CPS 3 7 CPS.

3 Turn table -axis laser range finder 6) CPS 6 CPS 7 CPS Slit-like laser 4 Measured local shape Measured local shape Measured local shape 3 3 Coordinate transformation by CPS data z y x Parent robot r ψ φ (a) Measurement by LRF (b) Positioning of child robots Aligned model Child robots () Robot and move. () Robot measures the position of robot. r ψ φ (3) Robot measures the position of robot. 5 r ψ φ, CPS (4) Robot moves and measures the position of robots and. ICP(Iterative Point) 3 Closest r ψ φ (c) Positioning of parent robot (d) Measurement by LRF 6 CPS 3 7 CPS..3% 6)6) AP-L TOPCON GPT-95A TOPCON () LMS SICK LMS5 SICK () A5 TOPCON 8mm A3 TOPCON (3) mm 3 cfl Information Processing Society of Japan

TOPCON) 95A, 5 Resolution SIG Technical Report IPSJ 6) AS- Auto-leveling system MD-9-TS (AP-L) Applied Geomechanics, Inc.

Leveling Max. range Before improvement ±4 ffi Read pitch and yaw angles from external attitude sensor 5 (GPT- Max.

elevation angles, and distance from total station Data correction Read pitch and yaw angles from internal attitude sensor

4 TOPCON) 95A, 5 Resolution SIG Technical Report IPSJ 6) AS- Auto-leveling system MD-9-TS (AP-L) Applied Geomechanics, Inc., MD-9-TS, 9, 3 AS- 8,9 4 GPT-95A 5 Total station 4 9 Two-axes attitude sensor (MD-9-TS) (AS-) (AS-, Rizumu) accuracy ± Leveling Max. range Before improvement ±4 ffi Read pitch and yaw angles from external attitude sensor 5 (GPT- Max. range ±6 After improvement Make total station level by auto-leveling system Automatic leveling device Read azimth, elevation angles, and distance from total station Data correction Read pitch and yaw angles from internal attitude sensor Read azimth, elevation angles, and distance from total station 8 Laser scanner 3. Data correction cfl Information Processing Society of Japan

5 m mm x 93mm y 5mm z 9mm (8.9m).54%.3% 6) 6) Error at the corner of the door is 98mm 3.34% 3m Error.6m y z x 35m m 5m mm x -47mm y 7mm z -78mm 7m Loop y x RGB-D 4 RGB-D 3 3 RGB-D 4. 6, RGB-D TOPCON A7P, 4 5 cfl Information Processing Society of Japan

3Hz sensor range :m 3:5m Depth (7.6, 7.7, 4.9) z [m] 5 4 3 (5.9, 37.7,.4) (3., -5., 4.9) 4 3 Start & Goal (-4.

6 3Hz sensor range :m 3:5m Depth (7.6, 7.7, 4.9) z [m] (5.9, 37.7,.4) (3., -5., 4.9) 4 3 Start & Goal (-4.4, -.3,.) y [m] - - x [m] 5 7 Microsoft Kinect 4 RGB-D Microsoft, Kinect, 6. Kinect RGB-D Microsoft Kinect field of view 57 ffi Horizontal field of view 43 ffi Vertical size 64 Λ 48 Image Scanning frequency Microsoft Kinect 7.3. Kinect ) 4.., CPS 3 6 cfl Information Processing Society of Japan

Levoy, M., Pulli, K., Curless, B., Rusinkiewicz, S., Koller, D., Pereira, L., Ginzton, ) M., Anderson, S., Davis, J., Ginsberg, J., Shade, J. and Fulk, D.

and Masuda, ) Bayon DigitalArchivalProject, In Proceedings of the Tenth International Con- T.: on Virtual Systemand Multimedia, pp.334 343 (4). ference Ishikawa, K., ichi Takiguchi, J., Amano, Y.

, Thayer, S., Whittaker, W., Baker, C., Burgard, W., Ferguson, D., 5) D., Montemerlo, M., Morris, A.C., Omohundro, Z., Reverte, C. and Whittaker, Haehnel, W. R.L.

7 Levoy, M., Pulli, K., Curless, B., Rusinkiewicz, S., Koller, D., Pereira, L., Ginzton, ) M., Anderson, S., Davis, J., Ginsberg, J., Shade, J. and Fulk, D.: The Digital Project: 3D Scanning of Large Statues, Proc. ACM SIGGRAPH, Michelangelo (). pp.3 44 Ikeuchi, K., Hasegawa, K., Nakazawa, A., Takamatsu, J., Oishi, T. and Masuda, ) Bayon DigitalArchivalProject, In Proceedings of the Tenth International Con- T.: on Virtual Systemand Multimedia, pp (4). ference Ishikawa, K., ichi Takiguchi, J., Amano, Y., Hashizume, T. and Fujishima, T.: Tun- 3) Measurement System Using a Mobile Mapping System, Journal nelcross-section Robotics and Mechatronics, Vol., No., pp (9). of (). pp.aa A8 Thrun, S., Thayer, S., Whittaker, W., Baker, C., Burgard, W., Ferguson, D., 5) D., Montemerlo, M., Morris, A.C., Omohundro, Z., Reverte, C. and Whittaker, Haehnel, W. R.L.: Autonomous exploration and mapping of abandoned mines, IEEE - Vol.5, No.8, pp (7). Tobata, Y., Kurazume, R., Noda, Y., Lingemann, K., YumiIwashita and Hasegawa, 7) Laser-based geometrical modeling of large-scale architecturalstructures using T.: multiple robots, Autonomous Robots, Vol.3, No. (). co-operative Kurazume, R., Noda, Y., Tobata, Y., Lingemann, K., Iwashita, Y. and Hasegawa, 8) Laser-based Geometric Modeling using Cooperative Multiple Mobile Robots, T.: IEEE International Conference onrobotics and Automation, pp.3 35 Proc. No.6, pp (995). Vol.3, Nüchter, A., Surmann, H., Lingemann, K., Hertzberg, J. and Thrun, S.: 6D SLAM ) an Application in Autonomous Mine Mapping, Proc. IEEE International Conference with onrobotics and Automation, pp (4). Weingarten, J. and Siegwart, R.: EKF-based 3D SLAM for Structured Environment ) Reconstruction, Proc. IEEE/RSJ International Conference on Intelligent and System,, pp (5). Robots Cole, D.M. and Newman, P.M.: Using Laser Range Data for 3D SLAM in Outoor ) Total station Corner mirror Parent robot 8.3%.5% RGB-D Microsoft Kinect JST (B) ) and Automation Magazine, Vol., No., pp.79 9 (4). Robotics CPS-SLAM - 6) (9). 9) cfl Information Processing Society of Japan

8 Proc. IEEE International Conference on Robotics and Automation, Environment, (6). pp Grisetti, G., Stachniss, C. and Burgard, W.: Improved techniques for grid mapping 3) rao-blackwellized particle filters, IEEE Transactions on Robotics, Vol.3, p. with (7). 7 Olson, E., Leonard, J. and Teller, S.: Fast iterative alignment of pose graphs with 4) estimates, Proc. IEEE International Conference onrobotics and Automation, poor (6). pp.6 69 Kümmerle, R., Steder, B., Dornhege, C., Ruhnke, M., Grisetti, G., Stachniss, C. 5) Kleiner, A.: On Measuring the Accuracy of SLAM Algorithms, Autonomous and Vol.7, No.4, pp (9). Robots, - 8 CPS-III 6) (998). - pp Kinect 7) 9 pp.o3 4 (). 8 cfl Information Processing Society of Japan

458 31,..., [ 123 ].,,, ( 1) : 2 A B, A,. 2,, B. A, B, 2,, A,, [ 425 ]. X A, A x ), B. 2. XO Y, xay, x, y B, x = AB cos, y = AB sin. (1) : A A

458 31,..., [ 123 ].,,, ( 1) : 2 A B, A,. 2,, B. A, B, 2,, A,, [ 425 ]. X A, A x ), B. 2. XO Y, xay, x, y B, x = AB cos, y = AB sin. (1) : A A 31 4 Vol. 31. 4 2010 4 Journal of Harbin Engineering University Ap r. 2010 doi: 10. 3969 / j. issn. 1006-7043. 2010. 04. 009,,, (,, 150001) :,. 2,, CCD,.,, RGB H IS. 3,,,.,,. : ; ; ; : O34 : A : 100627043