Research on Conductive Rubber Array Flexible Tactile Sensor Based on CNN-LSTM-Attention Hybrid Model

Flexible tactile sensors have attracted significant attention in the fields of industrial robotics and healthcare. However, the fabrication of low-cost, high-stability tactile sensors remains challenging. This study proposes a tactile sensor based on a conductive rubber array, which is composed of a conductive rubber array embedded with an intermediate support layer, encapsulated by flexible printed circuits (FPCs). The sensor employs a 4×4 layout (16 channels in total) to capture time-series voltage signals as inputs for a deep neural networe. Furthermore, using a convolutional neural network (CNN) + long short-term memory network (LSTM) + Attention hybrid model, the tactile sensing array achieves a classification accuracy of 97.2 % across diverse objects. This demonstrates promising application potential in complex scenarios and fulfills the integration and scalability requirements for robotic electronic skins.