Abstract
In daily contexts, fabrics embodied in garments are in contact with human body all the time. Since fabric material properties-such as softness or fineness-can be easily sensed by human fingertips, fabric materials can be roughly identified by fingertip sliding. Identification by simply touching and sliding is convenient and fast, although the room for error is always very large. In this study, a highly discernible fabric humanoid identification method with a fingertip structure inspired tactile sensor is designed to investigate the fabric material properties by characterizing the power spectrum integral of vibration signal basing on fast Fourier transform integral S(FFT), which is generated from a steel ball probe rubbing against a fabric surface at an increasing sliding velocity and normal load, respectively. kv and kw are defined as the slope values to identify the fabric surface roughness and hardness. A sample of 21 pieces of fabric categorized by yarn weight, weave pattern, and material were tested by this method. It was proved that the proposed humanoid sensing method has more efficient compared with fingertip sliding while it is also much more accurate for fabric material identification. Our study would be discussed in light of textile design and has a great number of potential applications in humanoid tactile perception technology.