Force and Thermal Sensing with a Fabric-based Skin

We present a novel fabric-based multimodal tactile sensing skin with three sensing modalities: force, actively heated temperature sensors to measure heat transfer and passive, unheated temperature sensors. In our evaluation, the skin recognized two materials during pressing and sliding tasks. Our method thermally distinguished pine wood from aluminum after detecting the contact force. With a support vector machine (SVM) classifier trained on 0.25 s of data, our method achieved a recognition accuracy of 96% for the pressing task and 73% for the sliding task.