Deep Learning for Control of Digital Systems
| Autor: | Ghadeer I. Maki, Zahir M. Hussain |
|---|---|
| Rok vydání: | 2021 |
| Předmět: | |
| Zdroj: | Journal of Physics: Conference Series. 1804:012086 |
| ISSN: | 1742-6596 1742-6588 |
| Popis: | Classically industrial systems apply a number of techniques to control their components, including the control system, which modify the relationship between input and output signals to configure the system to provide the required response. In most practical systems these signals are continuous, hence it is important to convert them into digital signals to be processed by digital systems. Despite the great development in technology, given the importance of the control system in relation to dynamic systems to achieve optimal performance, but classical control suffers from some important problems. The complexity of the control system represented by the program implementation algorithms and the loss of most information during the process of converting the system to digital and not adapting to external variables or with new updates. In this research classical control is replaced by deep neural networks, which is a thriving field with practical and medical applications and is characterized by its ability to learn and train as it is a branch of machine learning and artificial intelligence. The results proved that the functioning of the neural networks and their performance is similar to classical control systems, with the advantage of simplicity and adaptability. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|