RSM-Based Data-Driven Optimized Design of a 3-D-Printed Building Block-Type FBG Tactile Sensor for Nursing Robots

Abstract

This paper presents a three-dimensional printed building block-type tactile sensor based on fiber Bragg grating for force feedback of nursing robots with the achievement of shape estimation and pulse signal detection. The tactile sensor includes four sensing units and splicing grooves, which can be expanded like the building block to achieve large-area tactile perception. Each sensing unit is composed of a three-dimensional printed bellows elastomer and a suspended optical fiber inscribed with a fiber Bragg grating. A response surface method driven by simulation data has been applied to establish the stiffhess model of the sensing unit. The experimental sensitivity of the tactile sensor with optimization by the genetic algorithm is 225.038 pm/N. The response time to step excitation is 2 ms, and the signal-to-noise ratio of the measurement signal with a constant force of 0.255 N is 32.93 dB. A spliced tactile array assembled by two tactile sensors can effectively sense the targefs shape. The experimental errors of the tactile array for measuring the pulse rates of three testers are less than 4.4 %. The tactile array can effectively detect the pulse waveforms of males and females as well as before and after doing exercise. The respiratory signal has been obtained by fitting the peak value of the pulse signal with cubic spline interpolation. Robot-assisted pulse rate testing further validates the performances of the tactile array.