Abstract
Large area tactile sensing is widely needed in human machine interfaces, prosthetics and for collaborative robots. However, scaling tactile sensing to larger areas or to higher spatial resolution typically involves a direct tradeoff with the number of sensing units, wiring complexity, and fabrication cost. Inspired by the properties of human skin, super-resolution (SR) sensing has reduced the number of required sensors, but simultaneously achieving large area coverage with high spatial resolution, measured by the SR factor, remains unresolved. Here, we present a new SR method using a 2D meandering spoof surface plasmon polariton (SSPP) transmission line that solves all the challenges above. Combining principles of frequency-diverse computational sensing with deep neural networks, the proposed electromagnetic skin uses a single wire to achieve SR factors exceeding $10,000$, an order of magnitude more than other approaches, while covering sensing areas of $2500-28,800$ mm$^2$. The EM-skin facilitates human computer interaction with dynamic tactile stimulus and object recognition with high accuracies. Overall, this work provides a promising route towards highly scalable large-area e-skin with super-resolution sensing.