Learning-Based Approach for Non-Linear Applied Control
Rishabh Dev Yadav (Ph.D. Student)
Non-linear control systems are widely used in many real-world applications, such as robotics, aerospace, and power systems. Despite their widespread use, these systems often face significant challenges in achieving accurate and reliable control performance. This is due to the complex and unpredictable nature of non-linear systems, which can result in unpredictable behavior and instability. To overcome these challenges, researchers have developed various control methods, such as model-based adaptive and robust control. While these methods have demonstrated some success in controlling non-linear systems, they often require precise mathematical models and require significant expertise to design and implement. In addition, these methods are not always robust to changes in system dynamics and disturbances. In recent years, there has been increasing interest in using learning-based approaches for non-linear control. These methods leverage the power of machine learning algorithms to learn the dynamics of non-linear systems and generate optimal control actions. Unlike model-based methods, learning-based approaches do not require precise mathematical models, making them more flexible and adaptable to changes in system dynamics. However, the performance and limitations of these algorithms in non-linear control systems are not well-understood. The findings of this research will provide valuable insights into the use of learning algorithms for non-linear control, and will contribute to the development of more effective and efficient control strategies for non-linear systems. This will enable researchers and engineers to improve the performance and stability of these systems, leading to more efficient and reliable operations in a variety of applications including robotics, aerospace and mechatronics.
| Primary Host: | Wei Pan (University of Manchester) |
| Exchange Host: | Sihao Sun (Delft University of Technology) |
| PhD Duration: | 01 October 2023 - 31 December 2027 |
| Exchange Duration: | 01 January 2025 - 31 December 2025 - Ongoing |
