Gait Sensors with Customized Protruding Structures for Quadruped Robot Applications

Stable data acquisition and accurate recognition of motion states are critical for biomimetic robots operating in complex environments. This study proposes flexible gait sensors that can detect pressure and vibration for quadruped robots. These sensors are fabricated using a template-confined electrospinning technique, allowing for direct customization of protruding structures. The developed gait sensor exhibits a maximum capacitive sensitivity of 1.237 kPa^-1, a detection extending range up to 1000 kPa, and a fast response time of 5 ms. Leveraging their lightweight nature, these sensors can detect vibrations at various weight loads, frequencies, and amplitudes. Moreover, a recognition process combining these gait sensors with deep learning techniques for quadruped robot applications has been studied. It demonstrates the capability of the sensors to monitor diverse locomotion poses and states of the robot, achieving impressive accuracies of up to 97.50% for gait recognition and 98.04% for abnormal disturbances. This research offers potential applications in developing electronic skins for robots and provides promising solutions for enhancing robot performance in challenging environments.