A Flexible Artificial Sensory Nerve Enabled by Nanoparticle‐Assembled Synaptic Devices for Neuromorphic Tactile Recognition. Issue 24 (9th June 2022)

• Record Type: Journal Article
• Authors: Jiang, Chengpeng; Liu, Jiaqi; Yang, Lu; Gong, Jiangdong; Wei, Huanhuan; Xu, Wentao
• Is Part Of: Advanced science. Volume 9:Issue 24(2022)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2022-06-09
• Subjects: flexible synaptic electronics -- nanomaterial self‐assembly -- neuromorphic perception -- robotics and smart interfaces -- tactile recognition; Science -- Periodicals; 505
• DOI: 10.1002/advs.202106124
• Languages: English
• ISSNs: 2198-3844

Abstract

Abstract: Tactile perception enabled by somatosensory system in human is essential for dexterous tool usage, communication, and interaction. Imparting tactile recognition functions to advanced robots and interactive systems can potentially improve their cognition and intelligence. Here, a flexible artificial sensory nerve that mimics the tactile sensing, neural coding, and synaptic processing functions in human sensory nerve is developed to achieve neuromorphic tactile recognition at device level without relying on algorithms or computing resources. An interfacial self‐assembly technique, which produces uniform and defect‐less thin film of semiconductor nanoparticles on arbitrary substrates, is employed to prepare the flexible synaptic device. The neural facilitation and sensory memory characteristics of the proton‐gating synaptic device enable this system to identify material hardness during robotic grasping and recognize tapping patterns during tactile interaction in a continuous, real‐time, high‐accuracy manner, demonstrating neuromorphic intelligence and recognition capabilities. This artificial sensory nerve produced in wearable and portable form can be readily integrated with advanced robots and smart human–machine interfaces for implementing human‐like tactile cognition in neuromorphic electronics toward robotic and wearable applications. Abstract : A flexible artificial sensory nerve that achieves neuromorphic tactile recognition in a real‐time and near‐sensor manner is developed. In this system, tactile signals coded as spike trains are processed by synaptic devices with sensory memory, and tactile recognitions are performed at device level by directly using device output without relying on computing resources or algorithms.