E-skin for human–robot interaction: tactile sensing and edge computing architectures

Electronic skin (E-skin) has been a key enabling technology for robots that need to work safely, intelligently, and in physical proximity with humans. The domain has progressed from fingertip tactile arrays to large-area, full-body systems that can now sense pressure, shear, high-frequency vibration, proximity (i.e., the distinct sensation of “touching” another body), temperature, and even social touch. In this article, we showcase the architectural bottlenecks, material science challenges, and embedded computing requirements that influence whether E-skin can transition from laboratory prototypes to deployable robotic systems. From our analysis, we showcase that none of the currently available transduction principles satisfy all requirements of spatial resolution, conformability, low latency, low cost, calibration stability, and mechanical robustness. Capacitive and time-of-flight methods improve pre-touch awareness, pneumatic and barometric schemes enhance coverage and low cost, and tomographic approaches and row–column strategies reduce wiring, while deep learning techniques enable semantic recognition of human tactile gestures. Simultaneously, this paper claims that the next E-skin trends will be built on top of heterogeneous architectures composed of soft materials, distributed sensor nodes, deterministic safety loops, and edge computing platforms based on Arduino, Raspberry Pi, Onion Omega, and NVIDIA Jetson devices alike. Lastly, two comparative tables and a scoping process informed by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) were used to clarify the evidence base and technological trade-offs.