Change of Orientation of the TCP on using SearchL along the path

soumith0602

We are working on the Yumi robot with a yumipy library developed on python. The library uses MoveL and SearchL commands to move linearly from the starting position to the To position.
As we have observed, the orientation changes on the path from the initial position to the final position. The pose at the final position is the same as that of initial position but that is not the case in the linear path followed. However, on using the jogging control with the joystick on the IRC5 controller, the orientation remains the same all the way.

Does someone have any idea on how to implement the SearchL command in a similar way to have same orientation at all the stop points along the path?
Also, any idea on which command the jogging controller uses on the RAPID Code to move linearly?

graemepaulin

With a MoveL the TCP will wove in a straight line between the two points, and the orientation will be changed in equal steps from the start to the end points orientation.

So in your case (where the orientation is unchanged) there should be no re-orientation between the two points.

soumith0602

graemepaulin said:

With a MoveL the TCP will wove in a straight line between the two points, and the orientation will be changed in equal steps from the start to the end points orientation.

So in you case (where the orientation is unchanged) there should be no re-orientation between the two points.

Thanks for your reply, In our case when use the SearchL, the orientation did change while the robotic arm is moving, but we need to keep the same orientation as the start point even we interrupt the movement. I checked the Technical reference manual again and only the command - (CirPathMode - Tool reorientation during circle path) relates to the reorientation but ours is a linear movement. Or should I set the v_ori to zero in speeddata to prevent the orientation? What is the correct way to stop re-orientation during the movement from ToPoint to the TargetPoint by using the SearchL?

graemepaulin

How much movement are you seeing?

soumith0602

graemepaulin said:

How much movement are you seeing?

We aligned the right arm so the Euler Angle are 0, 0, 180, and the Target is 15 cm far in z direction. If we stop the movement, there usually are tiny changes less than 1 degree in EY and EX. Since the robot arm holds an camera, the small rotation can cause an obvious inclined view, which is not required.

graemepaulin

Why are you searching if you are also using the camera?

soumith0602

Well, we are using the SearchL command to stop at any position along the path from the initial to the final position. The camera is just for viewing the direction in which the robot is moving and in assisting the user.

graemepaulin

I think the level of movement you are seeing is just the fact that the robot motion computer is interpolating 7 axis to keep the TCP and orientation correct while moving along the path.

The manual does not give any figures for the orientation but the linear path accuracy is 1.3 mm (repeatability 0.1 mm), so there is going to be some orientation movement along the path.

If you want to correct the orientationn after you stop you can use RAPID to take the current position, and then make the orientation equal to your wanted values and move to this new corrected position.