| The continuous progress of information technologies has resulted in a notable evolution in human-machine interfaces,artificial perceptual systems and bio-robots.With the increase of data volume,traditional sensing devices based on von Neumann architecture face the bottleneck of power consumption and energy efficiency.Besides,traditional sensing devices based on silicon substrate are usually rigid and have poor compatibility with organisms in physical form,which cannot adapt to the development trend of intelligent wearable electronic systems in the future.Benefit from the highly parallel network structure of the biological neural network and its event-triggered computing mode,the biological sensory system realizes the perception and processing of external signals at a lower cost of space,time and power,and realizes the function of information processing and memory by synaptic weight modulation.Emerging flexible artificial sensory systems containing synaptic devices have been considered as a promising solution for processing massive data with low power consumption.In this thesis,we focus on the research of flexible neuromorphic-based artificial sensory devices and systems from the perspective of bio-inspired information perception,transmission,processing,and feedback.The main research results are as follows.1.Inspired by the biological tactile functions,we design a piezo-resistive tactile sensor using the single-wall carbon nanotubes and micro-pyramid-structured polyimide.The device has a high sensitivity with high stability by optimizing the micro-pyramid-structures size and film thickness.2.Inspired by the living creatures’ pain perception process,we constructed a flexible double-layer memristor using semiconducting single-wall carbon nanotubes covered by LiClO4 doped polyoxyethylene oxide as the channel materials.The conductance change of the device can be modulated by the intercalation of Li+in the carbon nanotubes under different voltages,which can mimic the functions of the neural system such as the sensation of pain,neuroprotection,and possible injuries.3.Inspired by the signal processing process of neurons and synapses,we demonstrated the artificial synapse that can achieve the neural-like signal processing to the voltage and light.The synaptic plasticity can be controlled by regulating protons,carriers,and defects in the device.Functions of neurons and synapses such as short-term potentiation(STP),long-term potentiation(LTP),and learning and forgetting behaviors can be mimicked and achieved.4.Inspired by biological sensation,perception,and action processes,we constructed an artificial somatosensory system with tactile sensation,neuromorphic processing,and instant feedback capabilities.The artificial system can achieve the spatial information-processing function for tactile stimulus.Also,the artificial muscle in the system can be triggered judgment circuits when the pressure reaches a threshold. |
