About the project
This PhD project aims to advance iterative learning control (ILC) by overcoming the limitation of requiring identical references in each trial, a constraint that can be impractical for many real-world applications. By combining optimization, machine learning, and biological insights, this research will develop ILC algorithms that learn faster and adjust to trial-varying tasks.
Iterative Learning Control (ILC) is a powerful tool for systems engaged in repetitive tasks, where performance improves by learning from previous attempts. However, conventional ILC designs assume a fixed reference trajectory across all trials, restricting their use in dynamic, real-world applications where task requirements may vary.
This PhD project seeks to eliminate this limitation by creating ILC algorithms that not only learn more efficiently but can also accommodate different reference trajectories in each trial, enhancing flexibility and robustness in diverse applications.
The research will develop innovative frameworks that fuse optimization and machine learning with principles inspired by biological learning processes. Optimization techniques will fine-tune control strategies, machine learning will introduce adaptability across task variations, and biological insights will inspire approaches to achieve faster, more efficient learning. Together, these elements will contribute to new algorithms that can generalize across similar yet distinct tasks, allowing systems to learn beyond simple repetition.
The successful candidate will focus on formulating new ILC methods, implementing these algorithms in simulations, and testing their effectiveness in real-world scenarios requiring task flexibility. The outcomes of this project are expected to have a transformative impact on fields such as robotics, autonomous systems, and manufacturing, where versatile and efficient learning algorithms are essential for enhanced performance, reduced downtime, and increased adaptability.