https://doi.org/10.1140/epjs/s11734-025-02081-1
Regular Article
Full-hardware design and application of memristor neural networks based on Widrow–Hoff algorithm circuit
1
The Key Laboratory of Micro-nano Sensing and IoT of Wenzhou, Wenzhou Institute of Hangzhou Dianzi University, 325038, Wenzhou, Zhejiang, China
2
The School of Electronics and Information, Hangzhou Dianzi University, 310018, Hangzhou, Zhejiang, China
3
Department of Integrated Circuits, Sun Yat-sen University, 518107, Shenzhen, Guangdong, China
4
Department of Electrical and Electronic Engineering, The University of Western Australia, Crawley, 6009, Perth, WA, Australia
a
youyuan-0213@163.com
b
luciali@ieee.org
Received:
31
July
2025
Accepted:
24
November
2025
Published online:
1
December
2025
This paper presents a design of Widrow–Hoff algorithm circuit, which, in combination with memristor synapse circuits and neuron circuits, enables a fully hardware-implemented supervised neural network. The Widrow–Hoff algorithm circuit functions as a backpropagation algorithm circuit, updating the synaptic weights within the hardware neural network. The memristor synapse circuit is responsible for realizing the multiplication of input signals and weights, while the neuron circuit performs the summation of weighted signals and applies the activation function to generate the output. Finally, the proposed circuit is applied to character recognition, achieving a fully hardware neural network capable of supervised learning for character identification. All circuits were simulated and verified using LTspice. This study provides both theoretical and experimental foundations for the full-circuit implementation of highly complex neural networks. Specifically, the fully hardware-based weight update method proposed in this paper can overcome the von Neumann bottleneck, effectively reducing the frequent migration of data and providing an effective reference for neural acceleration. Compared with hardware-software co-design approaches, this method offers significant advantages.
Copyright comment Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
© The Author(s), under exclusive licence to EDP Sciences, Springer-Verlag GmbH Germany, part of Springer Nature 2025
Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
