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

Size: px

Start display at page:

Download "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"

褰缇皮
5 years ago
Views:

1 31 4 Vol Journal of Harbin Engineering University Ap r doi: / j. issn ,,, (,, ) :,. 2,, CCD,.,, RGB H IS. 3,,,.,,. : ; ; ; : O34 : A : (2010) M utual localization of mobile robots using an omn i2vision system LU Jun, SUN L ing2li, MU Hai2jun, ZHU Q i2dan (College of Automation, Harbin Engineering University, Harbin , China) Abstract:A method for localization, the difficult p roblem of determ ining a robot s pose, was developed based on panoram ic vision. It solves the p roblem by using op tical sensor data shared among mobile robots. A testing system was developed with two mobile robots, each of which was equipped with a panoram ic CCD camera and a cylinder rendered in three colors for identifying the robot. Panoram ic im ages were obtained using a hyperbolic m irror based om ni2vision sensor. Through binarization, noise reduction and a region grow ing algorithm, feature regions were found. Then the regions were transformed from RGB color mode into H IS color mode. Based on the order of a robot cylinder s three colors, the robots were distinguished from the background and each robot identified. Then using the mutual localization algorithm, the coordinates of the robot in the world coordinate system were calculated. Finally, the experimental results were exam ined and sources of errors analyzed. The experimental results indicate that this localization m ethod for mobile robots, based on a panoram ic vision system, p roduced satisfactory results, and these methods have significant value for p ractical app lications. Keywords: omni2vision; mobile robot; mutual localization; im age p rocessing,.,. :. : 1). : : (19692),,, E2mail: lujun0260@ sina. com.,.. 2). : 1).,. 2).,,.,. 3).

458 31,..., [ 123 ].,,,. 1 1. 1 ( 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 B.

2 458 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 B. B X B Y B = x cos + y sin + X A, = - x sin + y cos + Y A. (2) (1) (2), 2 AB, B. 2 1 Fig. 1 The p rincip le figure of robots mutual localization based on panoram ic vision 1. 2,,,,. A (X A, Y A ), ( Fig. 2 The geometrical p rincip le of robots localization AB,, B (A ) 2. 3 B A. O 1 O 2 A B, R, 12 A B ( ). AB = R / sin ( 12 ). (3) 2

4, : 459 B A, 2 AB = R / sin ( 21 2 ) (4) 3 B A Fig. 3 The overlooking geometrical figure from robot B to robot A 1. 2. 2 21 12 21 12,. 4. : ( )., 360. : 360 ( ), 360 ( )., 21 12.

3 4, : 459 B A, 2 AB = R / sin ( 21 2 ) (4) 3 B A Fig. 3 The overlooking geometrical figure from robot B to robot A ,. 4. : ( )., 360. : 360 ( ), 360 ( )., , n, [m, m + n - 1 ], s t, = s + t - 2m (5) 2n B, 12 = t - s (6) n A, 21 = t - s (7) n 1. 3,.. : 1 : A A B ; 2 : B B A ; : X A B ; x A : A ; x B : B ; : B, ; 5, 2, 5 = , - 1, 1 > ; 1. (8) 4 Fig. 4 The cylinder expanding figure of the omni image

4 B, 5, 6,. 1) 0 <, = 2 + -, 2 + +, - 2 < 2 ; 0 2 < -. 5 Fig. 5 The figures for solving angle (9) 2) < 2, = , 2 + -, 3-2 < 2 ; 0 2 < -. (10)

5 4, : B Fig. 6 The figures for solving robot B s DOA angle 2,,,,. [ 728 ].,,,., 3,. RGB, HSI, I,, RGB H IS,, 3,,, [ 627 ]. 7, 8.

6 Fig. 7 The flowchart of target recognization ( a) ( b) ( c) 8 Fig. 8 The images of the original omni image and recognized image

4, : 463 3 1. B, A. ( B ) : R = 6. 5 cm, A, 2 AB, 2 ( 2 ), 1., 2,. 2. 2,,, 2.,, 2. 3. 9, A B, B, B A, A. : 1) AB = 98. 6 cm (20 ), 1 = 13. 68 ; 2) AB = 94.

5 118. 3 127. 4 138. 0 /cm ( ) 12. 2 18. 5 18. 3 27. 4 28.

7 4, : B, A. ( B ) : R = 6. 5 cm, A, 2 AB, 2 ( 2 ), 1., 2,. 2. 2,,, 2.,, , A B, B, B A, A. : 1) AB = cm (20 ), 1 = ; 2) AB = cm ( 21 ), 2 = ( 800 ) 1 1 AB Table 1 The com par ison between actua l da ta and ca lcula tion s for AB in exper im en t /cm /cm /cm ( ) AB Table 2 The com par ison between actua l da ta and ca lcula tion s for AB in exper inen t /cm ( /cm /cm 75 /cm ) ( a) A 4 ( b) B 9 Fig. 9 The expanding images of omni images that robot and get,,, 2 : 1). 2). 2. A, (1) (2), B 2

8 : 2 AB ;, ( 8) (10), B 1 2., B AB 1 2. (3) (4), AB AB : AB = AB AB 2 - R 2. (11) 2R : AB AB, R AB, (11) AB = AB 2 2R. (12) (12),, AB R ; R, AB AB.,, 1 2. : 1),. 2),,. 3). 4). : AB, AB, AB,., 2 [ 8 ]. 5,,.,.,,,.,.,. : [ 1 ],,. [ J ]., 2007, 33 (8) : YUAN Kui, L I Yuan, FANG L ixin. Reseach development situation for muti2robot[ J ]. Acta Automatica Sinica, 2007, 33 (8) : [ 2 ],,. [ J ]., 2007, 3: XU Guobao, YIN Yixin, ZHOU Meijuan. Technology situa2 tion and future on Intelligent mobile robot technology [ J ]. Robot Technique and App lication, 2007, 3: [ 3 ]BEOM H R, CHO H S. Mobile robot localization using a single rotating sonar and two passive cylindrical beacons [ J ]. Robotics, 1995, 13: [ 4 ] CHEN S. Quicktime VR an image2based app roach to virtual environment navigation: computer Graphics [ C ] / / Proc of SIGGRAPH 95. [ s. l. ], [ 5 ] SIEGWART R. NOURBAKHSH R, Autonomous mobile ro2 bots [M ].,. :, 2006: [ 6 ]. [D ]. :, 2006: MU Haijun. The localization research of the mobile robot based on the panoram ic vision system [ D ]. Harbin: Harbin Engineering University, 2006: [ 7 ],. [ J ]., 2004, 26 (1) : CH I J iannan, XU Xinhe. Research on simultaneous locali2 zation and mapp ing of mobile robot [ J ]. Robot, 2004, 26 (1) : [ 8 ]JENSFELT P, CGRUSTEBSEB H. Laser based position and tracking in an indoor environment[ C ] / / Proceedings of In2 ternational Symposium on Robotics and Automation. [ s. l. ], [ : ]

Microsoft Word - KSAE06-S0262.doc

Microsoft Word - KSAE06-S0262.doc Stereo Vision based Forward Collision Warning and Avoidance System Yunhee LeeByungjoo KimHogi JungPaljoo Yoon Central R&D Center, MANDO Corporation, 413-5, Gomae-Ri, Gibeung-Eub, Youngin-Si, Kyonggi-Do,