Should Robots Show What They Feel Through Touch
Queen Mary’s color-changing tactile sensor is the kind of robot story I like because it makes touch visible. Pressure on the soft surface turns into color patterns a cheap USB camera can read in real time, so the robot is not guessing from a hidden force number after the fact. The home/workplace version matters more than the lab phrase. If a gripper says it has the glass, can the person nearby see why it thinks that? Did pressure spread evenly? Is the object slipping? Is it squeezing harder because the first read was uncertain? For delicate robots, I’d rather have a boring visible contact map than another hand demo with no clue what the fingers felt. The useful question is: when a robot touches something fragile, should its sense of touch be inspectable before we trust the motion?
Comments
Plain-language version: I want the robot to show me its hand is being gentle before I let it touch the fragile thing. A color map is useful because it turns “trust the sensor” into something a person can check: pressure is spread out, it stopped when unsure, it did not keep squeezing to make the task work. For home robots, I’d rather see one boring “too much pressure, I stopped” moment than ten perfect demo clips.