Page 1 of 1

Uncertainty in the image data during each simulation step

Posted: 18 Jun 2019, 11:53
by denoza
Background:
I am trying to use the Vrep simulator to perform a visual servo control simulation using UR3 robot with eye in hand configuration.

Problem:
While using the B0 based remote api in Python, I noticed that at initial position of the robot upon starting and running the simulation, the image received is slightly different. This happens each time I run the simulation, the point to note is that this is a simple start and stop, I do not change the position of the robot. Since I want to use the image difference as the error therefore this uncertainty can cause problems during the control. In addition, if I run the simulation for the short time, this uncertainty can be big. Ideally the norm of image difference in float values should produce 0.0 (at initial position) but it can produce the values in the range 22.0 to 136.0. Is there a way to address this problem? Probably I am doing something wrong.

Configuration:
I am using the simpleTest.py script for simulation in synchronous mode. And my image size is (256,256).

Re: Uncertainty in the image data during each simulation step

Posted: 21 Jun 2019, 13:36
by coppelia
Hello,

I am curious whether you are deleting and constructing the b0RemoteApi object, between individual runs?

Cheers

Re: Uncertainty in the image data during each simulation step

Posted: 23 Mar 2023, 14:22
by SevenTRI
denoza wrote: 18 Jun 2019, 11:53 Background:
I am trying to use the Vrep simulator to perform a visual servo control simulation using UR3 robot with eye in hand configuration.

Problem:
While using the B0 based remote api in Python, I noticed that at initial position of the robot upon starting and running the simulation, the image received is slightly different. This happens each time I run the simulation, the point to note is that this is a simple start and stop, I do not change the position of the robot. Since I want to use the image difference as the error therefore this uncertainty can cause problems during the control. In addition, if I run the simulation for the short time, this uncertainty can be big. Ideally the norm of image difference in float values should produce 0.0 (at initial position) but it can produce the values in the range 22.0 to 136.0. Is there a way to address this problem? Probably I am doing something wrong.

Configuration:
I am using the simpleTest.py script for simulation in synchronous mode. And my image size is (256,256).
Excuse me, could you teach me how to get DH parameters and Jocabian of the robot arm? Thanks in advance.