The Impact of Lattice Structure and Skin on Performance of Magnetic Tactile Sensors

One way to give robots a sense of touch is a magnetic tactile sensor: a magnet sits inside a soft skin above a Hall-effect sensor, and pressing the skin moves the magnet and changes the field the sensor reads. Most studies change the magnet, sensor, and skin together, so the effect of the skin is hard to isolate. We asked how the internal structure and material of the skin change the sensor's sensitivity, repeatability, and hysteresis with everything else fixed. We hypothesized that skins that deform more under load would produce larger magnetic signals, so that cast silicone and open lattices would be the most sensitive and lattices enclosed by a solid wall would be the least. We simulated solid, honeycomb, and gyroid blocks in ANSYS, then built five skins (cast silicone and four 3D-printed TPU lattices) and loaded each from 0 to 80 g and back three times above an MLX90393 sensor. Silicone was the most sensitive skin (0.135 µT/g), followed by a 10% honeycomb (0.084 µT/g) and a 10% gyroid (0.063 µT/g), even though simulation predicted the gyroid would deform 22 times more than the honeycomb. An outer wall almost erased the response of a 5% gyroid, while a walled 5% hexagonal lattice gave the most repeatable readings. Skin design changed sensitivity more than tenfold, and the most sensitive skin was also the least repeatable, so the best skin depends on what the sensor is for.