A low-power and flexible bioinspired artificial sensory neuron capable of tactile perceptual and associative learning. Issue 7 (19th January 2023)

• Record Type: Journal Article
• Authors: Xia, Qing; Qin, Yuxiang; Zheng, Anbo; Qiu, Peilun
• Is Part Of: Journal of materials chemistry. Volume 11:Issue 7(2023)
• Page Start: 1469
• Page End: 1477
• Publication Date: 2023-01-19
• Subjects: Materials -- Periodicals; Chemistry, Analytic -- Periodicals; Biomedical materials -- Research -- Periodicals; 543.0284
• DOI: 10.1039/d2tb02408j
• Languages: English
• ISSNs: 2050-750X

Abstract

Abstract : A flexible and low energy consumption artificial tactile neuron was proposed and experimentally demonstrated by integrating a dual network hydrogel-based sensor and a low power memristor. Abstract : Biomimetic haptic neuron systems have received a lot of attention from the booming artificial intelligence industry for their wide applications in personal health monitoring, electronic skin, and human–machine interfaces. In this work, inspired by the human tactile afferent nerve, we developed a flexible and low energy consumption artificial tactile neuron, which was constructed by combining a dual network (DN) hydrogel-based sensor and a low power memristor. The tactile sensor (ITO/PAM:CS-Fe 3+ hydrogel/ITO) serves as E-skin, with mechanical properties including pressure and stretching. The memristor (Ti:ITO/BiFeO3 /ITO) serving as an artificial synapse has low power (∼3.96 × 10 −7 W), remarkable uniformity, a large memory window of 500 and excellent plasticity. Remarkably, the pattern recognition simulation based on a neuromorphic network is conducted with a high recognition accuracy of ∼89.81%. In the constructed system, the artificial synapse could be activated by the electrical information from the E-skin induced by an external pressure, to generate excitatory postsynaptic currents. The system shows functions of perception and memory functions, and it also enables tactile associative learning. The present work is important for the development of empowering robots and prostheses with the capability of perceptual learning, and it provides a paradigm for next-generation artificial sensory systems with low-power, wearable and low-cost features.