Smart and Low-Cost Flexible Strain Sensor Based on Graphene-MWCNT Porous Elastic Sponge for Home Control and Object Grasping Recognition Using Machine Learning

Flexible sensors are attracting significant interest due to their pivotal role in applications such as human activity monitoring and human–computer interaction. The low-cost polyurethane (PU) sponge with high elasticity and repeatability possesses significant potential for applications in flexible electronics and smart devices. Hence, we developed a piezoresistive pressure sensor based on a (graphene-MWCNT)/PU sponge composite, fabricated via a straightforward polymerization and dipping-drying process. This method leverages the stable porous structure of the PU sponge to ensure the robust adhesion of graphene and MWCNTs onto its skeleton, leading to the formation of an effective conductive network. The resulting low-cost sensor demonstrates excellent sensitivity (0.1 kPa^–1) and remarkable stability, maintaining its performance for 1000 cycles. Moreover, the smart sensor can be snugly affixed to the human body for the detection of human motion signals, enabling applications such as monitoring diverse human motions, recognizing different objects, and controlling external devices such as an LED light and a fan. Interestingly, by integrating the sensors into an array with a signal acquisition circuit, we developed a system capable of providing tactile feedback by mapping the real-time spatial pressure distribution during complex tasks. When combined with a deep learning algorithm, this system successfully classified five different grasped objects with an accuracy of 97.6%. These results highlight the significant potential of this sponge-based pressure sensor for applications in advanced household appliances and AI-integrated real-time control systems.