A FBG tactile sensor with rolling ball device for surface feature measurement and recognition

Touch is essential for acquiring surface characteristics, but material variations and environmental conditions such as temperature and humidity can affect surface friction, complicating texture recognition. This study proposes a rolling ball-type Fiber Bragg Grating (FBG) tactile sensor, which leverages FBG’s high sensitivity and electromagnetic interference resistance to detect textures through rolling, thereby minimizing friction-induced interference. Static calibration tests showed the sensor’s average sensitivity to be 20.8 pm/N (loading) and 20.5 pm/N (unloading), both with a linearity of 0.998, and low repeatability and hysteresis errors of 1.6 % and 1.1 %, respectively. The sensor was evaluated with concave-convex textures of 1.5 mm depth and spacings of 2, 3, and 4 mm, demonstrating that texture parameters could be accurately calculated with errors less than 0.1 mm across different conditions. Signal amplitude correlated with texture spacing when the contact force was consistent. A Back Propagation (BP) classification algorithm achieved 100 % accuracy at speeds of 5, 10, and 15 mm/s, with an overall accuracy of 98.52 %. This sensor provides valuable insights for advanced tactile perception in robotic fingertips.