# What are robot arm singularities in the context of the Space Shuttle and/or ISS?

This comment on this answer has got me baffled.

What are robot arm singularities, in the context of the Space Shuttle's or ISS's large robotic Canadian arms?

Are these singularities related to spacecraft attitude, or just to the mathematics behind the robotic articulation algorithms?

• Robot kinematic singularities (unrelated to spacecraft...they are just properties of controlling an articulated arm): quora.com/Robotics-What-is-meant-by-kinematic-singularity Sep 20, 2018 at 6:45
• @BlueCoder okay I see, thanks: robohub.org/… (and videos, and links therein) so perhaps this is not related to spacecraft attitude control, or only in a subtle way.
– uhoh
Sep 20, 2018 at 7:06
• A simple example of a singularity with a human arm: Point your thumb toward your eyes. Now do the same with your arm fully extended (elbow unbent). Unless you are extremely flexible, you cannot do so. Sep 20, 2018 at 11:28

A singularity in this context is simply a configuration of the manipulator system in which a degree of freedom is lost.

For example, for the Shuttle and ISS arms, the elbow singularity occurs when the elbow joint angle nears 180 degrees; the degree of freedom lost is the ability to translate the end effector away from the shoulder joint.

The relationship to the question about attitude control is that the software would have to avoid all these singularities if the arm were being used for attitude control. Just as they have to be avoided when the arm is being used to manipulate a payload.

Sometimes in the latter case the arm has to be reconfigured to avoid a singularity, and that sounds undesirable for the attitude control case.

Edit: list of singularities for the shuttle arm

1. Shoulder yaw; loss of +/- Y translation
2. Planar pitch; loss of translation away from the shoulder
3. Wrist yaw; loss of some roll commands
• This is a really insightful source, thanks! I'll continue to read today. I noticed, on page 28 of the slides (references) near the end: "Consequences of limited insight into proprietary model • 20 years of discussion of model/response differences at different forums without resolution • Operations designed conservatively to cover model uncertainties" I think having lessons learned sections is laudable.
– uhoh
Sep 27, 2018 at 2:10
• I'm glad you mentioned this, I didn't really read the paper once I found the picture I wanted. Looking at it now, it is pretty good. I taught many lessons in the facility shown in Figure 14. Sep 27, 2018 at 2:18