A voltage mode memristor bridge synaptic circuit with memristor emulators

Area of Science:

• Neuroscience
• Electrical Engineering
• Computer Science

Background:

• Previous work proposed a memristor bridge neural circuit for signed synaptic weighting, verified through software simulation.
• The study extends prior research by focusing on the physical circuit implementation of the memristor synapse.

Purpose of the Study:

• To advance the circuit implementation of a memristor bridge neural circuit capable of signed synaptic weighting.
• To construct and verify a simple neural network using memristor emulator circuits for synaptic weighting and summation.

Main Methods:

• The memristor bridge synapse architecture was built using memristor emulator circuits.
• A neural network was created by integrating these memristor emulator synapses with differential amplifier circuits.
• Circuit feasibility was confirmed using SPICE simulations.

Main Results:

• The memristor bridge synapse architecture was successfully implemented using emulator circuits.
• A functional neural network combining these synapses and differential amplifiers was constructed.
• SPICE simulations validated the operational feasibility of the proposed neural circuit.

Conclusions:

• The memristor bridge neural circuit, implemented with emulator circuits, is a feasible approach for signed synaptic weighting.
• This work demonstrates a practical step towards building neural networks with memristor-based synaptic components.
• The findings pave the way for more complex neuromorphic computing architectures.