A Dual‐Modal Programmable E‐Skin with Tunable In‐Sensor Touch Computing

ABSTRACT Touch perception endows embodied intelligence with autonomous cognition and decision‐making. However, conventional sensing frameworks based on von Neumann architecture suffer from sensing and computing separation, leading to data redundancy, complex parameter tuning, and limited scalability. Here, a dual‐modal programmable intelligent electronic skin (DPI e‐skin) is proposed unprecedentedly with touch event‐driven in‐sensor computing, enabling efficient position and pressure synchronous recognition. With an expandable multilayer stacking architecture, the DPI e‐skin reconstructed the signal form at the device level, directly outputting continuous position information and gradient pressure classification that matched the cognition scale. It significantly reduced redundant data and maintained high sensing performance with precise position resolution (<500 µm) and adjustable wide detection range (>380 kPa). Moreover, the pressure thresholds of different layers in the DPI e‐skin could be dynamically tuned to support diverse applications, like multi‐degree‐of‐freedom control and a typing system. Benefitting from the synergy of position and pressure, the DPI e‐skin serves as an encryption device to achieve accurate users’ static and dynamic feature cognition with high accuracy (>96%). This study provides an innovative design paradigm for high‐performance and deeply integrated touch perception that promotes the development of interactive intelligence, with the aim of achieving greater efficiency, security, and personalization.