sawyer Accuracy of limb data and data published via ROS

[Atmaraaj]

I am doing a research which requires accurate coordinate and joint values from Sawyer, and as such I have the following questions to ask about Sawyer's joint accuracy.

1. What are the accuracies of each limb length, joint angle values that are published via ROS and of the kinematic solvers?
    
2. Is it possible to directly get the poses (coordinates) of any joint of choice? For now, we just calculate these values from the limb lengths and joint values published, thus the first question.
    
3. Where is the virtual gripper/end-effector point relative to Sawyer's arm? As I understand, the end-effector point is needed to calculate the Inverse Kinematics, but I am unsure where exactly this point is.
    
4. If it is possible to obtain this end-effector position, how can I tune/calibrate this position when a different gripper is attached? I know that the Intera UI allows some configuration, but is this possible with Intera SDK too?

I would gladly explain more about my aims if more explanation is required. Thank you.

[Atmaraaj]

Any help, please?

[senaa]

Have you looked into whether the robot's tf data can give you what you need? 

 

[Atmaraaj]

I can read the data published, but I would like to know the accuracy of this data itself. Since Sawyer is not an industrial robot which usually has a much higher accuracy, I need to know of the uncertainties of the data published. Rethink Robotics, I assume, should have details on the accuracies.