Year: 2024
Type: conference-paper
Abstract: In the modern world, autonomous control systems have been applied in a wide range of fields, profoundly impacting everyone's life. They not only free humans from these tasks, but also provide a more efficient and reliable alternative beyond the limitations of humans. In recent years, the rapid development of artificial intelligence (AI) and machine learning (ML) has added wings to these systems. In this paper, we will focus on the robustness of autonomous control systems, which is important in real life but is often overlooked. We propose a framework combining pre-determined algorithms and several AI based neural networks including LSTM, MLP and DQN. Then we pick one of the circumstance, a post-disaster UAV rescue mission featuring measurement interference and potential breakdowns, and simulate it with a computer program. A specific system is built and trained based on our framework and evaluated in the environment. The results show that the system exhibits significantly superior robustness and control performance compared with other established models. Looking ahead, the potential scalability and applicability of this framework can be further explored and can serve as a reference in various fields. Our code can be downloaded at https://github.com/ZhenXiao-Jiang/robust_control_system.git/.
Author: Zhenxiao Jiang
Institution: Tongji University
Language: English
Subfield: Aerospace Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: gold
OpenAlex ID: W4409964101