Fiber Bragg grating-based touch-slip sensor for surface roughness detection.

To enable humanoid prosthetic hands to accurately identify and grasp objects, a touch-slip sensor based on fiber Bragg grating (FBG) was proposed in this paper. The sensor was designed with a double-layer sensing structure to detect three-dimensional force, sliding information, surface roughness, and compensate for ambient temperature using a reference grating. To analyze the relationship between the sensor surface structure parameters and FBG's vibration signal, the contact sliding model was introduced. By using a finite element simulation, the contact sliding model was validated, and the surface structure parameters of the sensor were determined. A series of experiments were conducted on the sensor's three-dimensional force detection, temperature calibration, sliding measurement, and surface roughness detection. Using machine learning methods, a regression prediction model was created. The sensor could detect the surface roughness on the sample plate surface more accurately with a maximum error of 8.58×10. This sensor has the following advantages: a simple structure, low cost, high linearity, and quick response time. It provides a new design solution for the detection of touch-slips on humanoid prosthetic hands.