How does a robot feel when a stopper comes loose?
Pulling out a stopper has a tiny turning point: it gives, and your hand changes what it's doing. I'd love a robot helping at a bench to notice that moment rather than keep tugging. The PanoTac research project gives robot fingertips a detailed map of pressure and sideways forces. Its test-tube experiment uses touch to detect stopper release during removal. The same project tests grain scooping and salt pouring; these are controlled research tasks, not evidence that the robot can run a lab. That little change in resistance makes touch easier to understand than a claim about humanlike hands. Seeing where the tube is doesn't tell the robot everything about whether the stopper is still holding on. For a future shared-bench trial, I'd want to try different stopper fits and see whether the robot can finish gently without a nearby person hovering to catch the tube. That's a trial I'd propose, not a result the project establishes. Which everyday job would you most want a robot to do by feel?
Comments
I'd separate a missed release from a false one. Continuing to pull after the stopper comes free and stopping while it is still stuck leave the person at the bench with different problems. PanoTac's page reports 90% success for test-tube preparation, which also includes reorientation and insertion; that isn't a separate accuracy figure for detecting release. In the proposed trial with different stopper fits, I'd record those two mistakes and whether anyone had to steady the tube or finish the removal. That would help a lab decide whether someone can step away from this chore, rather than just watch a robot do it.