Search

Advanced filters
Published After
Published Before

Search Results

Found one item.

  • INVESTIGATION OF SYNAPTIC PLASTICITY IN MEMRISTIVE CROSS-POINT STRUCTURES FOR NEUROMORPHIC ROBOTIC SYSTEMS

    R.V. Tominov , Z. Е. Vakulov , V.I. Varganov , I.О. Ignatieva , V. А. Smirnov
    200-207
    2025-12-30
    Abstract ▼

    The results show multilevel resistive switching and synaptic plasticity of a memristive cross-point based on a nanocrystalline zinc oxide film. It is shown that with a decrease in the amplitude and duration of input pulses, the memristive cross-point demonstrates resistive states from 4.27 × 105 Ohm to 8.34 × 107 Ohm. It is shown that the switching energy of some synaptic levels is picojoules, which is promising for creating compact low-power neuromorphic systems. Thus, it is shown that nanocrystalline ZnO films have synaptic plasticity, i.e. When applying voltage pulses, large limits and duration can vary depending on the synaptic levels.
    The fabricated memristive cross-point demonstrates paired-pulse facilitation PPF at tp from 1 ms to 10 ms and pair-pulse depression PPD at tp from 50 ms to 100 ms. The analysis of the experimental results of the PPF and PPD study showed that an increase in the number of pulses from 10 to 90 leads to an increase in postsynaptic current EPSC from 32 μA to 73 μA for tp = 1 ms, from 31 μA to 59 μA for tp = 5 ms, from 31 μA to 48 μA for
    tp = 10 ms, and a decrease in EPSC from 30 μA to 25 μA for tp = 50 ms, from 30 μA to 17 μA for tp = 70 ms, from 30 μA to 5 μA for tp = 100 ms. From the obtained results it follows that the interval between pulses, the higher the PPF index, thus it can be concluded that the manufactured memristive cross-point based on ZnO nanocrystalline films imitates the crucial plasticity of the biological synapse, in which the plasticity of PPF and PPD is determined by the concentration of Ca+ ions and which plays a role in many biological functions of the brain, such as determining the key source of sound, pattern recognition, associative learning, filtering unnecessary. information. The obtained results can be used for hardware implementation of neural networks, neuromorphic structures of robotic complexes, prostheses and artificial intelligence systems

1 - 1 of 1 items