Automating the Design of Scalable and Efficient IoT Architectures Using Generative Adversarial Networks and Model-Based Engineering for Industry 4.0
The increasing demand for efficient and scalable IoT systems in Industry 4.0 has driven the development of increasingly complex architectures. However, the design of these architectures presents significant challenges, particularly in dense networks where latency, energy consumption, and scalability...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11053865/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | The increasing demand for efficient and scalable IoT systems in Industry 4.0 has driven the development of increasingly complex architectures. However, the design of these architectures presents significant challenges, particularly in dense networks where latency, energy consumption, and scalability constraints affect operational performance. Traditional approaches, such as heuristic and genetic algorithms, have proven insufficient in automating and optimizing large-scale IoT configurations, resulting in a high design and validation time cost. This work proposes an innovative approach based on integrating generative adversarial networks (GANs), model-based engineering (MBSE), and MATLAB simulations to automate and optimize the design of IoT systems. Using GANs to generate scalable architectures and validate them using MBSE, our approach automates both the creation and optimization of configurations, ensuring compliance with technical and operational constraints. The results show that the proposed design achieves an average latency between 30 and 70 ms, improving by 10% compared to the reference values of the IoT Benchmark Dataset. Energy consumption remains between 3.2 W and 15.6 W, representing a 12% reduction compared to traditional systems. In addition, packet loss remains below 7%, exceeding the average of conventional systems. |
|---|---|
| ISSN: | 2169-3536 |