Repetitive and learning motion control in printing systems
Learning control can significantly improve the performance of motion systems, such as printing systems, by learning from past error signals. The aim of this project is the development of such learning controllers, with attention to design and implementation aspects.
Printing systems are subject to increasing requirements regarding accuracy, throughput, and cost. Traditional controllers are designed before commissioning the system, and remain fixed throughout the life-cycle of the system. The aim of this research is to develop and implement controllers that enhance performance by learning from past error signals. Indeed, many disturbances are highly repetitive, either due to batch-to-batch operation, and/or due to repeating movements of components, etc. These disturbances occur in many subcomponents of complex printing systems, which may exhibit a substantial amount of interaction. The aim of this research includes, but is not limited to
- characterization of repeating disturbances, both batch-to-batch and periodic disturbances;
- the development of a unified framework to attenuate these disturbances, i.e., comprising both traditional iterative learning control approaches and repetitive control approaches;
- the development of design approaches within this framework, which are particularly applicable to large scale multivariable systems, both in a decentralized and centralized way;
- appropriate incorporation of digital implementation aspects, thereby addressing the inherent trade-off between accuracy and equipment cost requirements; and
- development of a user-friendly toolbox with experimental validation of the developed approaches on a range of printing systems.