It has been interesting to discuss Collaborative Robot technologies with different audiences over the years. For those that haven’t had much experience with robot technology, the idea of a force limited robot certainly has a “wow” factor to it and normally there are very few technical concerns. It is very normal for there to be an expectation that if it is advanced enough to safely operate without a safety cage (pending a risk assessment), then it must be advanced enough to do virtually anything.
This assumption can often escalate to where even experienced users will ask if they can use collaborative robot where other robots have failed in the past. Sometimes the answer is yes, but unfortunately, as impressive as force limited robots are, they still experience a number of typical challenges seen by other robots. One of those challenges is something called a ‘Singularity’.
For those of you, who like me, grew up watching Star Trek, this has nothing to do with Quantum Singularities (I bet more than one of you thought we were going in that direction), this has to do with the challenges of trying to translate mathematical representations to the physical world. This is not unlike building an equation in an excel spreadsheet and getting the frustrating ‘#DIV/0’ error or, even worse, finding yourself with an infinite loop error.
Without going deep into the math, check out what our friends at Robotiq have published on the subject for a great visual of what singularities in a robot look like and how to work around them.
So as impressive as collaborative robots are, they are still limited by the mathematical models that are used to control the motion of all other robots. Can we someday come up with a robot that will figure how to avoid these singularities it’s own or at least warn us it’s own limitations? Yes, but we’ll leave Deep Learning and robots for another day.