A Digital-Analog Bimodal Memristor Based on CsPbBr₃ for Tactile Sensory Neuromorphic Computing

Area of Science:

• Materials Science
• Neuroscience
• Computer Engineering

Background:

• Memristors offer potential for advanced information processing but enabling bimodal digital and analog functions remains challenging.
• Fabricating artificial sensory neural network systems requires integrated components with biological synaptic functions.
• Developing efficient artificial synapses and sensory receptors is crucial for neuromorphic computing and human-machine interaction.

Purpose of the Study:

• To develop a CsPbBr₃-based memristor exhibiting digital and analog bipolar bimodal resistive switching.
• To construct an artificial sensory neural network system integrating the memristor with a tactile sensor array.
• To demonstrate the system's capability in mimicking biological synaptic functions and performing pattern recognition.

Main Methods:

• Fabrication of a CsPbBr₃-based memristor with high ON/OFF ratio, long retention, and stable endurance.
• Development of a 5x5 spinosum-structured piezoresistive sensor array for tactile sensing.
• Integration of the memristor and sensor array to create an artificial sensory neural network.
• Testing the system's performance in recognizing handwritten patterns using supervised learning.

Main Results:

• The CsPbBr₃ memristor demonstrated excellent performance characteristics including multilevel resistance memory.
• The piezoresistive sensor array exhibited high sensitivity, durability, and fast response time.
• The integrated system achieved a high accuracy of 94.44% in recognizing handwritten letters.
• The system successfully mimicked synaptic plasticity and neuromorphic computing functions.

Conclusions:

• The developed CsPbBr₃-based memristor effectively functions as an artificial synapse with bimodal capabilities.
• The artificial sensory neural network system shows significant potential for tactile sensing and pattern recognition.
• This research highlights the application of digital-analog bimodal memristors in human-machine interaction, prosthetics, and artificial intelligence.