Title: Deep learning-based controller design and performance enhancement of BLDC motor drive for electric vehicle technology

Authors: Anurag Singh; Shekhar Yadav; Sandesh Patel; Nitesh Tiwari

Addresses: Department of Electrical Engineering, Madan Mohan Malaviya University of Technology, Gorakhpur, UP, India ' Department of Electrical Engineering, Madan Mohan Malaviya University of Technology, Gorakhpur, UP, India ' Department of Electrical Engineering, Madan Mohan Malaviya University of Technology, Gorakhpur, UP, India ' Department of Electrical Engineering, Madan Mohan Malaviya University of Technology, Gorakhpur, UP, India

Abstract: BLDC motors are mostly used in industrial applications, particularly in electric vehicles (EVs). Nowadays, several studies are being conducted on BLDC motors due to the increasing demand for EVs. The speed performance of the current controller (CC)-based BLDC motor on PI, NNF, and CN networks analysis with the help of MATLAB and Simulink software at fixed and variable speeds. The CC CN provides satisfactory speed regulation in comparison to conventional PI and CC NNF controllers. The CC CN reduces settling time by approximately 20%, minimises peak overshoot by 15%, and enhances speed response by 25% compared to traditional PI controllers. The present investigation addresses the limitations, including excessive overshoot and sluggish response, of conventional PI controller-based CC solutions in BLDC motors. It suggests utilising NNF and CN networks to combine DL with CC to enhance speed performance, smoothness, and responsiveness.

Keywords: BLDC motor; PI controller; deep learning; current controller; neural-net fitting; custom network.

DOI: 10.1504/IJPT.2025.149491

International Journal of Powertrains, 2025 Vol.14 No.3, pp.222 - 239

Received: 15 Aug 2024
Accepted: 11 May 2025
Published online: 04 Nov 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article