Skin-Inspired Composite Tactile Sensor with Equivalent Gradient Modulus for Grip Monitoring

Tactile sensing devices with skin-like perception capabilities are a prerequisite for the application of industrial robots in intelligent manufacturing. However, current tactile pressure sensors face challenges in simultaneously quantifying both static pressure and dynamic sliding stimuli with high accuracy, which hinders robotic arms from maintaining stable gripping operations. In this paper, a flexible composite tactile sensor with equivalent gradient modulus (EGM) is proposed. The high-modulus pyramidal structure enables efficient stress transmission and rapid concentration at its tip. Conversely, the low-modulus hole arrays accommodate sensitive layer deformation while providing superior pressure tolerance for the sensor. Tactile sensor with crack-blocking sensing plane achieves high sensitivity to microstimuli. The optimized sensor exhibits high sensitivity, wide detection range (0.085 kPa–1 at 0–160 kPa; 0.368 kPa–1 at 160–300 kPa) and fast response (80 ms/96 ms, response/recovery time). Meanwhile, the EGM tactile sensor accurately distinguishes point-contact pressure and area (98.6% for 60 pressure–area combinations), while identifying sliding texture/contact area (86.4% for 5 objects). Incorporating a deep separable convolutional network with channel attention mechanism achieves an efficient balance between recognition accuracy (95.70%) and parameter optimization in classifying 12 grasp states, providing a promising solution for stable gripping in industrial automation.