Model design and exponential state estimation for discrete-time delayed memristive spiking neural P systems.

Area of Science:

• Computational Neuroscience
• Artificial Intelligence
• Materials Science

Background:

• Spiking Neural P Systems (SNPS) provide computational support for neural morphology and AI chips, offering high performance.
• Memristors, as emerging devices, uniquely integrate memory and computation, making them suitable for SNPS synapses.
• Existing SNPS often use resistors, limiting integration potential with advanced memory-based computing elements.

Purpose of the Study:

• To design and model a discrete-time memristive spiking neural P system (MSNPS) by replacing resistors with memristors.
• To analyze the impact of time delays and discretization on the MSNPS.
• To develop sufficient conditions for achieving exponential state estimation in the proposed MSNPS.

Main Methods:

• Circuit design integrating memristors into the SNPS framework.
• Mathematical modeling of the MSNPS based on the circuit design.
• Analysis of time delays and discretization effects.
• Application of Lyapunov functional theory for state estimation.

Main Results:

• Successful construction of a discrete-time MSNPS mathematical model.
• Establishment of sufficient conditions for exponential state estimation.
• Validation of the theoretical findings through a numerical simulation example.

Conclusions:

• The proposed MSNPS offers a novel architecture for AI chips by leveraging memristor properties.
• The derived conditions ensure reliable exponential state estimation for the MSNPS.
• This work paves the way for more efficient and integrated neuromorphic computing systems.