Neural Networks in Electrical and Computer Engineering: A Comprehensive Review
DOI:
https://doi.org/10.65204/djes.v3i2.547Keywords:
Neural networks, deep learning, electrical engineering, computer engineering, power systems, embedded systems, communication networks, neuromorphic computing.Abstract
Artificial Neural Networks (ANNs) have evolved into the core technology of electrical, computer, and industrial engineering, providing potent tools for intelligent modelling, optimization, prediction, as well as autonomous decision-making. Their ability to estimate very nonlinear and dynamic systems with data-driven learning and adaptability has allowed engineers to overcome most of the limitations that are related to the traditional analytical and mathematical methods. With modern engineering systems growing more complex due to the emergence of smart grids, interconnected embedded devices, industrial automation and extensive communication networks, the scope of neural networks (NNs) has grown exponentially.
This review gives a detailed and structured review of the technologies of the NN in terms of engineering. It starts with the theoretical background of neural computation, learning algorithms, and activation functions and then goes further by exploring some of the most significant architectures, including multilayer perceptrons (MLPs), convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory networks (LSTMs), and advanced deep learning models. The review identifies the use of these architectures in different engineering fields, such as stability assessment of power systems, prediction of loads and renewable energy, fault detection in electrical machine, signal and image processing, embedded AI in microcontrollers, robotic control of industries, and intelligent wireless communication systems.
Moreover, the current issues that ANN deployment faces, including the requirement of large annotated datasets, computational intensity, model interpretability, training stability, energy consumption, and hardware implementation limits in embedded and real-time applications are discussed in the paper. New research areas, including neuromorphic hardware, edge-AI acceleration, hybrid neuro-fuzzy system, reinforcement learning-based control and AI-controlled autonomous industrial production are also discussed to indicate future directions.
In general, this review is expected to represent a comprehensive reference to the researchers, engineers, and practitioners who would like to integrate NNs into the contemporary electrical, computer, and industrial engineering systems and provide an insight on the potential challenges, opportunities, and promising directions of the further developments.
References
Y. LeCun, Y. Bengio, and G. Hinton, “Deep learning,” Nature, vol. 521, pp. 436–444, 2015.
S. Haykin, Neural Networks and Learning Machines, 3rd ed. Pearson Education, 2009.
I. Goodfellow, Y. Bengio, and A. Courville, Deep Learning. MIT Press, 2016.
M. Hossain and H. Mohamed, “Machine learning based intelligent control for smart grids: A comprehensive survey,” IEEE Access, vol. 8, pp. 150110–150138, 2020.
S. J. Kazmi, A. Mateen, and H. Farooq, “Energy forecasting using deep neural networks for power systems,” Electric Power Systems Research, vol. 189, 2020.
A. S. Abdelhady, “Artificial intelligence techniques for power system protection,” IET Generation, Transmission & Distribution, vol. 14, no. 17, pp. 3362–3373, 2020.
A. G. Hussien and M. El-Wakad, “Fault detection and classification in transmission lines using neural networks,” IEEE Transactions on Power Delivery, vol. 36, no. 5, 2021.
A. Krizhevsky, I. Sutskever, and G. Hinton, “ImageNet classification using deep convolutional neural networks,” Communications of the ACM, vol. 60, no. 6, pp. 84–90, 2017.
D. Silver et al., “Mastering the game of Go with deep neural networks and tree search,” Nature, vol. 529, pp. 484–489, 2016.
M. Abadi et al., “TensorFlow: Large-scale machine learning on heterogeneous systems,” 2016.
G. Hinton et al., “Deep neural networks for acoustic modeling in speech recognition,” IEEE Signal Processing Magazine, vol. 29, no. 6, pp. 82–97, 2012.
. Lee, H. Kao, and S. Yang, “Industrial AI: Artificial intelligence for industry 4.0 systems,” Manufacturing Letters, vol. 18, pp. 20–23, 2018.
L. Ren et al., “Deep learning-based fault diagnosis in manufacturing systems,” IEEE Transactions on Industrial Electronics, vol. 66, no. 4, 2019.
Z. Zhang et al., “Deep learning for intelligent manufacturing: A review,” Journal of Manufacturing Systems, vol. 57, pp. 127–143, 2020.
K. J. Hunt, D. Sbarbaro, R. Zbikowski, and P. Gawthrop, “Neural networks for control systems—A survey,” Automatica, vol. 28, no. 6, pp. 1083–1112, 1992.
S. Levine, C. Finn, T. Darrell, and P. Abbeel, “End-to-end training of deep visuomotor policies,” Journal of Machine Learning Research, vol. 17, pp. 1–40, 2016.
M. H. Wang et al., “Neural network-based adaptive control for nonlinear robotic systems,” IEEE Transactions on Neural Networks and Learning Systems, vol. 31, 2020.
H. Ye, G. Li, and B. Juang, “Power of deep learning for channel estimation and signal detection,” IEEE Communications Magazine, vol. 57, no. 3, 2019.
S. Han, H. Mao, and W. Dally, “Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding,” International Conference on Learning Representations (ICLR), 2016.
M. Kim et al., “Microcontroller-based embedded neural networks using optimized lightweight models,” IEEE Internet of Things Journal, 2022.
G. Huang et al., “Edge AI: On-device intelligence for next-generation engineering systems,” IEEE Transactions on Industrial Informatics, vol. 19, 2023.
F. Bianchi et al., “Neural networks for smart grid automation: Trends and challenges,” Energy Reports, vol. 9, 2023.
X. Chen et al., “AI-enabled predictive maintenance using deep learning in industrial systems,” Expert Systems with Applications, vol. 228, 2024.
A. Das et al., “Neuromorphic computing for efficient AI hardware,” IEEE Transactions on Neural Networks and Learning Systems, 2024.
