Е-Публикации
Технически университет - София

Abstract: Memristors are state-of-the-art nano-scale electronic components characterized by excellent switching and memorizing capabilities, small energy usage and a good compatibility to CMOS integrated circuits. They are potentially applicable in neural nets, memory crossbars, and different electronic schemes. This paper suggests a plain LTSPICE model of a modified transfer function and a neuron, based on memristors. In neurons’ implementation, synapses are realized with single memristors, leading to a significantly reduced circuit complexity. The scaling and summing circuits are built on utilization of memristors and op-amps. The offered modified tangent-sigmoidal transfer function is realized with memristors and MOS transistors. For the analyses, a simple and high-speed memristor model is proposed. The offered memristor-based neuron is analyzed in LTSPICE and MATLAB. A comparison of the obtained results approves the proper action of the suggested neuron. The realized memristor neuron is an essential step towards engineering of complex neural nets and implementation in ultra-high-density integrated neural chips.

References

1. Aggarwal, C., “Neural Networks and Deep Learning,” Springer Int. Publ. AG: Germany, Berlin, 2018; ISBN 978-3-319-94463-0, p. 497.
2. Kim, P., 2017. “Matlab deep learning. With machine learning, neural networks and artificial intelligence” APress, ISBN 978-14842-28456.
3. Adhikari, S.P., Yang, C., Kim, H. and Chua, L.O.,. “Memristor bridge synapse-based neural network and its learning,” IEEE Trans. neural networks and learning systems, 2012, vol. 23, issue 9, pp.1426-1435.
4. Xu, W., Wang, J. Yan, X.,. “Advances in memristor-based neural networks,” Frontiers in Nanotechnology, 2021, vol. 3, pp. 1-14.
5. Strukov, D.B., Snider, G.S., Stewart, D.R., Williams, S., „The missing memristor found,“Nature 2008, vol. 453, pp. 80 – 83.
6. Ascoli, A., Tetzlaff, R., Biolek, Z., Kolka, Z., Biolkova, V., Biolek, D. “The Art of Finding Accurate Memristor Model Solutions,” IEEE Journal of Emerg. Sel. Top. Circ. Syst., 2015, vol. 5, pp. 133–142.
7. Lehtonen, E., Laiho, M., “CNN using memristors for neighborhood connections,” In Proc. Conf. CNNA, USA, 2010, pp. 1 – 4.
8. Z. Kolka, V. Biolkova and D. Biolek, "Simplified SPICE model of TiO2 Memristor,” Int. Conf. MEMRISYS, Cyprus, 2015, pp. 1-2.
9. S. Dautovic, N. Samardzic, A. Juhas, A. Ascoli, R. Tetzlaff, "Simscape and LTspice models of HP ideal generic memristor based on finite closed form solution for window functions," 2021 ICECS, pp. 1-6.
10. Campbell, K.A., 2017. “Self-directed channel memristor for high temperature operation,” Microelectronics journal, 59, pp.10-14.
11. Wen, S., Xie, X., Yan, Z., Huang, T., Zeng, Z., 2018. “General memristor with applications in multilayer neural networks,” Neural Networks, 103, pp.142-149.
12. Su, B., Cai, J., Wang, Z., Chu, J., Z., Y., “A π-Type Memristor Synapse and Neuron With Structural Plasticity,” Fr. Ph., 2022, pp.1-11.
13. Hong, Q., Zhao, L., Wang, X.,. “Novel circuit designs of memristor synapse and neuron,” Neurocomputing, 2019, vol. 330, pp.11-16.
14. Wang, S., Song, L., Chen, W., Wang, G., Hao, E., Li, C., Hu, Y., Pan, Y., Nathan, A., Hu, G., Gao, S., 2023. “Memristor-Based Intelligent Human-Like Neural Computing, Adv. El. Mat., 9, 1, pp. 1 – 39.
15. Mladenov, V., "A Unified and Open LTSPICE Memristor Model Library," MDPI Electronics, 2021, vol. 10, no. 13.
16. V. Mladenov, Georgi Tsenov, Stoyan Kirilov, "LTSPICE Memristor Neuron with MOS Transistor-Based Logarithmic-Sigmoidal Activation Function," acc. publ. 18th CNNA, https://cnna.duth.gr/.
17. Solovyeva, E. B., Azarov, V. A., "Comparative Analysis of Memristor Models with a Window Function Described in LTspice," El. Con. Rus, 2021, pp. 1097-1101.
18. May, C., "Passive Circuit Analysis with LTspice® - An Interactive Approach," Springer, 2020, ISBN 978-3-030-38304-6, pp. 763.

Issue