Date of Award

1-7-2021

Publication Type

Doctoral Thesis

Degree Name

Ph.D.

Department

Electrical and Computer Engineering

First Advisor

Majid Ahmadi

Second Advisor

Arash Ahmadi

Rights

info:eu-repo/semantics/embargoedAccess

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.

Abstract

Memristor is considered as a suitable alternative solution to resolve the scaling limitation of CMOS technology. In recent years, the use of memristors in circuits design has rapidly increased and attracted researcher’s interest. Advances have been made to both size and complexity of memristor designs. The development of CMOS transistors shows major concerns, such as, increased leakage power, reduced reliability, and high fabrication cost. These factors have affected chip manufacturing process and functionality severely. Therefore, the demand for new devices is increasing. Memristor, is considered as one of the key element in memory and information processing design due to its small size, long-term data storage, low power, and CMOS compatibility. The main objective in this research is to design memristor-based arithmetic circuits and to overcome some of the Memristor based logic design issues. In this thesis, a fast, low area and low power hybrid CMOS memristor based digital circuit design were implemented. Small and large-scale memristor based digital circuits are implemented and provided a solutions for overcoming the memristor degradation and fan-out challenges. As an example, a 4- bit LFSR has been implemented by using MRL scheme with 64 CMOS devices and 64 memristors. The proposed design is more efficient in terms of the area when compared with CMOS- based LFSR circuits. The simulation results proves the functionality of the design. This approach presents acceptable speed in comparison with CMOS-based design and it is faster than IMPLY-based memrisitive LFSR. The propped LFSR has 841 ps de-lay. Furthermore, the proposed design has a significant power reduction of over 66% less than CMOS-based approach. This thesis proposes implementation of memristive 2-D median filter and extends previously published works on memristive Filter design to include this emerging technology characteristics in image processing. The proposed circuit was designed based on Pt/TaOx/Ta redox-based device and Memristor Ratioed Logic (MRL). The proposed filter is designed in Cadence and the memristive median approved tested circuit is translated to Verilog-XL as a behavioral model. Different 512 _ 512 pixels input images contain salt and pepper noise with various noise density ratios are applied to the proposed median filter and the design successfully has substantially removed the noise. The implementation results in comparison with the conventional filters, it gives better Peak Signal to Noise Ratio (PSNR) and Mean Absolute Error (MAE) for different images with different noise density ratios while it saves more area as compared to CMOS-based design. This dissertation proposes a comprehensive framework for design, mapping and synthesis of large-scale memristor-CMOS circuits. This framework provides a synthesis approach that can be applied to all memristor-based digital logic designs. In particular, it is a proposal for a characterization methodology of memristor-based logic cells to generate a standard cell library for large scale simulation. The proposed framework is implemented in the Cadence Virtuoso schematic-level environment and was veri_ed with Verilog-XL, MATLAB, and the Electronic Design Automation (EDA) Synopses compiler after being translated to the behavioral level. The proposed method can be applied to implement any digital logic design. The frame work is deployed for design of the memristor-based parallel 8-bit adder/subtractor and a 2-D memristive-based median filter.

Available for download on Friday, January 07, 2022

Share

COinS