Comparative Analysis of Reconfigurable Platforms for Memristor Emulation

Area of Science:

• Electronics
• Materials Science
• Computer Engineering

Background:

• The memristor, the fourth fundamental electronic element, possesses unique memory and resistance properties.
• Current memristor research is limited by the lack of direct electronic solutions, relying on numerical models or emulators.
• Memristor emulators are crucial for studying and designing memristor-based systems.

Purpose of the Study:

• To propose and compare two reconfigurable hardware architectures for prototyping non-linear memristor emulators.
• To evaluate the suitability of Field-Programmable Analog Arrays (FPAA) and Field-Programmable Gate Arrays (FPGA) for memristor emulation.
• To provide a detailed comparison of FPAA and FPGA for memristor emulator implementation.

Main Methods:

• Proposed two distinct reconfigurable hardware architectures: FPAA and FPGA.
• Implemented non-linear memristor emulators on both FPAA and FPGA platforms.
• Conducted a comparative analysis of the FPAA and FPGA approaches for memristor emulation.

Main Results:

• Both FPAA and FPGA platforms successfully enabled the prototyping of non-linear memristor emulators.
• FPAA offers easy programming and reprogramming capabilities.
• FPGA provides high performance, area density, and parallelism for memristor emulation.

Conclusions:

• FPAA and FPGA are viable and effective platforms for developing memristor emulators.
• The choice between FPAA and FPGA depends on specific application requirements regarding flexibility versus performance.
• These hardware platforms can be integrated into more complex systems like neural networks and CNNs.