The research of memristor-based neural network components operation accuracy in control and communication systems

Danilin, S.N.; Shchanikov, S.A.; Galushkin, A.I.

doi:10.1109/sibcon.2015.7147034

Cited by 20 publications

(11 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Additionally, in order to take the practical imperfections of physically implemented systems (e.g. uncertainty in programming RRAM devices [35]) we have used previously developed techniques for keeping practical errors within acceptable tolerance levels [51,52].…”

Section: Machine Learning Set-upmentioning

confidence: 99%

Designing a bidirectional, adaptive neural interface incorporating machine learning capabilities and memristor-enhanced hardware

Shchanikov

Zuev

Bordanov

et al. 2021

Chaos, Solitons & Fractals

View full text Add to dashboard Cite

Section: Machine Learning Set-upmentioning

confidence: 99%

Designing a bidirectional, adaptive neural interface incorporating machine learning capabilities and memristor-enhanced hardware

Shchanikov

Zuev

Bordanov

et al. 2021

Chaos, Solitons & Fractals

View full text Add to dashboard Cite

“…For example, Danilin et al in [21] aimed to create a general approach to developing methods and algorithms designed for defining and providing memristor-based artificial neural network (ANNM) components operation accuracy in control and communication systems. They list factors that affect the operation accuracy of memristors used as the synapses in the ANNM.…”

Section: Related Workmentioning

confidence: 99%

Modelling Oil Pipelines Grid: Neuro-fuzzy Supervision System

Saeed¹,

Abbod²

2017

IJISA

View full text Add to dashboard Cite

Abstract-One of the major challenges for researchers and governments across the world is reducing resourceswaste or loss. Resources loss can happen if there is not a capable control system that contributes to environmental change. The specific aim is to create user-friendly control and monitoring system to reduce the waste in resources. New Artificial intelligence techniques have been introduced to play an important part in developing such systems.In oilfields, the oil is extracted then distributed via oil pipes until it reaches the end consumer. This operation will occur without a full and complete monitoring for the oil in the pipeline's journey to the provider. Although, the existing oilfield monitoring systems can communicate locally but they will not send information back to the main provider. That means the provider is not aware of the whole circumstances happened in the transportation process. That gives the provider no control on the process. For example, a sudden decision from the main provider to stop transporting to a specific destination or knowing where the leakage is and which pipe is leaking in the pipelines grid. This paper, introduces for the first-time oilfield pipeline Neuro-fuzzy (NF) supervision system using Simscape simulation software package. This system can be the first step solution to keep real time communication between the main provider and the oil transportation process in the oilfields and enables the provider to have full supervision on the oil pipes grid. The simulation supervision system illustrates a clear real-time oilfield pipeline grid that gives the provider the ability to control and monitor pipeline grid and prioritise the recovering process. The two parameters selected for control and monitoring were volume and pressure. The results in this paper show full control for the NF supervision system on the transportation process.

show abstract

“…In addition to the aforementioned reasons, the authors of the present paper have investigated and demonstrated in a number of research reports (Danilin et al, 2015(Danilin et al, , 2016 that the ANNM should be studied, designed, fabricated and operated as integrated physical and informational systems implemented with trainable hardware and software.…”

Section: Introductionmentioning

confidence: 99%

“…Many fault models of digital circuit elements used for hardware implementation of ANN and methods of their accounting are considered in the paper (Torres-Huitzil and Girau, 2017). They can be used in the design of digital parts of ANN, but at the same time, one of the most important unsolved problems in this area is the development of algorithms and approaches to determine and ensure the required operation accuracy of ANNM implemented on analog circuits elements, as well as determine and provide their fault tolerance (FT) and reliability (Danilin et al, 2015(Danilin et al, , 2016. At the informational level, FT is associated with the sensitivity of ANN, the analysis methods for which are given in the paper (Yeung et al, 2010).…”

Section: Introductionmentioning

confidence: 99%

Design and Hardware Implementation of Memristor-Based Multilayer Perceptron Network for a Bidirectional Adaptive Neural Interface

Shchanikov

Zuev

Bordanov

et al. 2021

2021 Third International Conference Neurotechnologies and Neurointerfaces (CNN)

View full text Add to dashboard Cite

A key problem at hardware implementation of artificial neural networks based on memristors (ANNM) is to ensure the required accuracy of their operation at the transition from models to real fabricated memristive devices. Due to a number of factors, such as the imperfections in state-of-theart memristors and memristive arrays, ANNM design and tuning methods, additional computation errors occur during the process of ANNM hardware implementation. This article proposes a general approach to the simulation and design of a multilayer perceptron (MLP) network on the basis of cross-bar arrays of metal-oxide memristive devices. The proposed approach uses the ANNM theory, tolerance theory, simulation methodology and experiment design. The tolerances analysis and synthesis process is performed for the ANNM hardware implementation on the basis of two arrays of memristive microdevices in the original 16×16 cross-bar topology being a component of bidirectional adaptive neural interface for automatic registration and stimulation of bioelectrical activity of a living neuronal culture used in robotics control system. The ANNM is trained for solving a nonlinear classification problem of stable information characteristics registered in the culture grown on a multi-electrode array. Memristive devices are fabricated on the basis of a newly engineered Au/Ta/ZrO 2 (Y)/Ta 2 O 5 /TiN/Ti multilayer structure, which contains self-organized interface oxide layers, nanocrystals and is specially developed to obtain robust resistive switching with low variation of parameters. An array of memristive devices is mounted into a standard metalceramic package and can be easily integrated into the neurointerface circuit. Memristive devices demonstrate bipolar switching of anionic type between the high-resistance state and low-resistance state and can be programmed to set the intermediate resistive states with a desired accuracy. The ANNM tuning, testing and control are implemented by the FPGA-based control subsystem. All developed models and algorithms are implemented as Python-based software.

show abstract

The research of memristor-based neural network components operation accuracy in control and communication systems

Cited by 20 publications

References 7 publications

Designing a bidirectional, adaptive neural interface incorporating machine learning capabilities and memristor-enhanced hardware

Designing a bidirectional, adaptive neural interface incorporating machine learning capabilities and memristor-enhanced hardware

Modelling Oil Pipelines Grid: Neuro-fuzzy Supervision System

Design and Hardware Implementation of Memristor-Based Multilayer Perceptron Network for a Bidirectional Adaptive Neural Interface

Contact Info

Product

Resources

About