## simspark-devel

 [simspark-devel] The angle in the horizontal plane. From: feng - 2011-05-20 14:07:16 Attachments: Message as HTML Hi all,There is a simple problem that has been bothering me for some time.Form a flag,we can get the Distance , the Theta,and the Phi. The description of Theta in User's Manual is:the angle in the horizontal plane.But when i raise NAO's head a little,Theta increases.Here is my test (with agent stand on(-5.2,0,0),do nothing):Head2 Angle :30 angle to RFlag1 :27.76 angle to RFlag2 :-27.65Head2 Angle :0 angle to RFlag1 :24.11 angle to RFlag2 :-23.98Head2 Angle :-30 angle to RFlag1 :26.87 angle to RFlag2 :-27.04I am not clear on this problem,please help.Best regards,Feng
 Re: [simspark-devel] The angle in the horizontal plane. From: Drew Noakes - 2011-05-20 14:14:00 Attachments: Message as HTML Hi Feng, Have you read the Wiki article on vision perceptors?  It has quite a lot of information that should help. http://simspark.sourceforge.net/wiki/index.php/Perceptors#Vision_Perceptors If it's still unclear, then please indicate how the document could be improved. Drew. ________________________________ From: feng To: simspark-devel@... Sent: Friday, 20 May 2011, 15:07 Subject: [simspark-devel] The angle in the horizontal plane. Hi all, There is a simple problem that has been bothering me for some time. Form a flag,we can get the Distance , the Theta,and the Phi. The description of Theta in User's Manual is:the angle in the horizontal plane.But when i raise NAO's head a little,Theta increases. Here is my test (with agent stand on(-5.2,0,0),do nothing): Head2 Angle :30 angle to RFlag1 :27.76 angle to RFlag2 :-27.65 Head2 Angle :0 angle to RFlag1 :24.11 angle to RFlag2 :-23.98 Head2 Angle :-30 angle to RFlag1 :26.87 angle to RFlag2 :-27.04 I am not clear on this problem,please help. Best regards, Feng ------------------------------------------------------------------------------ What Every C/C++ and Fortran developer Should Know! Read this article and learn how Intel has extended the reach of its next-generation tools to help Windows* and Linux* C/C++ and Fortran developers boost performance applications - including clusters. http://p.sf.net/sfu/intel-dev2devmay _______________________________________________ Simspark Generic Physical MAS Simulator simspark-devel mailing list simspark-devel@... https://lists.sourceforge.net/lists/listinfo/simspark-devel
 Re: [simspark-devel] The angle in the horizontal plane. From: Stefan Glaser - 2011-05-20 16:20:31 Attachments: Message as HTML Hi Feng, I will check the description of the wiki once again according to our agent-implementation. But since I did this already several times, I hope there is no error left any more. Keep in mind, that the camera is located in the center of NAO's head - but the pitch axis of the head is located a bit below. Since your changes are relatively small, this could already be an indicator. On the other hand, you could also try to look sidewards and check the second angle, which is provided to you. This way you can cancel out other values... Good luck, Stefan Am 20.05.2011 16:07, schrieb feng: > Hi all, > There is a simple problem that has been bothering me for some time. > Form a flag,we can get the Distance , the Theta,and the Phi. The description of Theta in User's Manual is:the angle in the horizontal plane.But when i raise NAO's head a little,Theta increases. > Here is my test (with agent stand on(-5.2,0,0),do nothing): > Head2 Angle :30 angle to RFlag1 :27.76 angle to RFlag2 :-27.65 > Head2 Angle :0 angle to RFlag1 :24.11 angle to RFlag2 :-23.98 > Head2 Angle :-30 angle to RFlag1 :26.87 angle to RFlag2 :-27.04 > I am not clear on this problem,please help. > Best regards, > Feng > > > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > > > _______________________________________________ > Simspark Generic Physical MAS Simulator > simspark-devel mailing list > simspark-devel@... > https://lists.sourceforge.net/lists/listinfo/simspark-devel
 Re: [simspark-devel] The angle in the horizontal plane. From: Stefan Glaser - 2011-05-20 20:37:54 Attachments: Message as HTML Hi Feng, I had a second look at your problem. As I wrote, you have a small gap between the camera and the pitch axis of the head, but if it would be this gap, then the resulting change should look different. I reproduced your situation on my machine and I was able to record similar values. Here my results: where: - Vec is the perceived polar coordinates in cartesian cordinates - res the resulting vision vector after an inverse rotatation of the movement - pos the resulting position calculation based on the known position of the flags NECK PITCH CASE ============================================================ (n hj2) (ax -30.00)) (See (G2R (pol 15.74 -4.50 30.91)) (G1R (pol 15.72 4.22 30.75)) (F1R (pol 17.20 26.69 25.43)) (F2R (pol 17.20 -26.85 25.41)) Vec1: (13.46, -1.05, 8.08) res: (15.70, -1.05, 0.27) pos: (-5.20, 0.01, 0.52) Vec2: (13.47, 0.99, 8.03) res: (15.68, 0.99, 0.22) pos: (-5.18, 0.05, 0.57) Vec3: (13.87, 6.97, 7.38) res: (15.71, 6.97, -0.54) pos: (-5.21, 0.02, 0.54) Vec4: (13.86, -7.01, 7.38) res: (15.69, -7.01, -0.53) pos: (-5.19, 0.01, 0.53) (HJ (n hj2) (ax -0.00)) (See (G2R (pol 15.74 -3.90 0.94)) (G1R (pol 15.74 3.77 0.94)) (F1R (pol 17.21 24.09 -1.70)) (F2R (pol 17.19 -24.15 -1.84)) Vec1: (15.70, -1.07, 0.25) res: (15.70, -1.07, 0.25) pos: (-5.20, 0.02, 0.54) Vec2: (15.70, 1.03, 0.25) res: (15.70, 1.03, 0.25) pos: (-5.20, 0.01, 0.54) Vec3: (15.70, 7.02, -0.51) res: (15.70, 7.02, -0.51) pos: (-5.20, -0.02, 0.51) Vec4: (15.67, -7.02, -0.55) res: (15.67, -7.02, -0.55) pos: (-5.17, 0.02, 0.55) (HJ (n hj2) (ax 30.00)) (See (G2R (pol 15.74 -4.41 -28.97)) (G1R (pol 15.73 4.23 -29.06)) (F1R (pol 17.19 27.58 -29.00)) (F2R (pol 17.23 -27.51 -28.70)) Vec1: (13.72, -1.05, -7.62) res: (15.70, -1.05, 0.26) pos: (-5.20, 0.00, 0.53) Vec2: (13.71, 1.01, -7.64) res: (15.69, 1.01, 0.23) pos: (-5.19, 0.03, 0.56) Vec3: (13.32, 6.96, -8.33) res: (15.70, 6.96, -0.55) pos: (-5.20, 0.03, 0.55) Vec4: (13.40, -6.98, -8.27) res: (15.74, -6.98, -0.46) pos: (-5.24, -0.01, 0.46) NECK YAW CASE ============================================================ (HJ (n hj1) (ax -30.00)) (See (G2R (pol 15.75 26.30 1.00)) (G1R (pol 15.71 33.57 0.69)) (F1R (pol 17.20 54.11 -1.79)) (F2R (pol 17.20 5.99 -1.83)) Vec1: (14.11, 6.97, 0.27) res: (15.71, -1.01, 0.27) pos: (-5.21, -0.03, 0.52) Vec2: (13.08, 8.68, 0.18) res: (15.67, 0.97, 0.18) pos: (-5.17, 0.07, 0.61) Vec3: (10.07, 13.92, -0.53) res: (15.69, 7.02, -0.53) pos: (-5.19, -0.02, 0.53) Vec4: (17.09, 1.79, -0.54) res: (15.70, -6.99, -0.54) pos: (-5.20, -0.00, 0.54) (HJ (n hj1) (ax 0.00)) (See (G2R (pol 15.75 -3.62 1.03)) (G1R (pol 15.75 4.00 1.01)) (F1R (pol 17.19 24.03 -1.88)) (F2R (pol 17.17 -24.15 -1.98)) Vec1: (15.71, -0.99, 0.28) res: (15.71, -0.99, 0.28) pos: (-5.21, -0.05, 0.51) Vec2: (15.70, 1.09, 0.27) res: (15.70, 1.09, 0.27) pos: (-5.20, -0.04, 0.52) Vec3: (15.69, 6.99, -0.56) res: (15.69, 6.99, -0.56) pos: (-5.19, 0.00, 0.56) Vec4: (15.65, -7.02, -0.59) res: (15.65, -7.02, -0.59) pos: (-5.15, 0.02, 0.59) (HJ (n hj1) (ax 30.00)) (See (G2R (pol 15.75 -33.51 1.06)) (G1R (pol 15.76 -25.84 1.08)) (F1R (pol 17.17 -6.05 -2.00)) (F2R (pol 17.20 -53.85 -1.79)) Vec1: (13.12, -8.69, 0.29) res: (15.71, -0.96, 0.29) pos: (-5.21, -0.08, 0.50) Vec2: (14.18, -6.86, 0.29) res: (15.71, 1.14, 0.29) pos: (-5.21, -0.09, 0.50) Vec3: (17.06, -1.80, -0.59) res: (15.68, 6.96, -0.59) pos: (-5.18, 0.03, 0.59) Vec4: (10.14, -13.88, -0.53) res: (15.72, -6.95, -0.53) pos: (-5.22, -0.04, 0.53) As you see, this effect is more related to pitch movements and the corder flags - the yaw case is not affected as much and as well not in the same way. This would also indicate a relation to the camera-pitch-gap, but if you look at the resulting positions, this effect isn't realy present any more. Overall I think it's related to the general noise of the vision, even though it looks dubious. Hope I could help you, Stefan Am 20.05.2011 16:07, schrieb feng: > Hi all, > There is a simple problem that has been bothering me for some time. > Form a flag,we can get the Distance , the Theta,and the Phi. The description of Theta in User's Manual is:the angle in the horizontal plane.But when i raise NAO's head a little,Theta increases. > Here is my test (with agent stand on(-5.2,0,0),do nothing): > Head2 Angle :30 angle to RFlag1 :27.76 angle to RFlag2 :-27.65 > Head2 Angle :0 angle to RFlag1 :24.11 angle to RFlag2 :-23.98 > Head2 Angle :-30 angle to RFlag1 :26.87 angle to RFlag2 :-27.04 > I am not clear on this problem,please help. > Best regards, > Feng > > > > ------------------------------------------------------------------------------ > What Every C/C++ and Fortran developer Should Know! > Read this article and learn how Intel has extended the reach of its > next-generation tools to help Windows* and Linux* C/C++ and Fortran > developers boost performance applications - including clusters. > http://p.sf.net/sfu/intel-dev2devmay > > > _______________________________________________ > Simspark Generic Physical MAS Simulator > simspark-devel mailing list > simspark-devel@... > https://lists.sourceforge.net/lists/listinfo/simspark-devel