Modern Linear Feedback Control in Mechatronics: Towards a Systematic Design Flow for Practicing Control Engineers

In a society that is depending on technology more than ever, scientists and engineers are continuously aiming at improving comfort, safety, energy efficiency, reliability, productivity, accuracy, etc. Today, one of the main drivers of persistent technological innovation is control or automation. Control systems have become omnipresent since the emergence of digital electronics in the 20th century, and it has even become difficult to imagine cars, computers, heating systems, household appliances, production machines, etc. that are not equipped with them. The oldest formal approaches to compose control algorithms focused on simple linear (electronic) filters such as PID controllers, which are -rather surprisingly- still widely recognized as the standard in industry. The sharp contrast with advanced control design methods that resulted from intensive academic research over the past decades is remarkable and raises questions about their practical significance. This work addresses exactly these questions, in particular for modern linear feedback control and with a strong focus on mechatronic applications. The contributions of the dissertation emphasize that the design procedure for linear feedback controllers is fairly systematic and is therefore ready to be widely adopted by practicing control engineers, provided that the adequate tools are available to support them.

On the one hand, the work reviews the fundamental concepts of loop shaping and robust control design for linear systems. Then, a solid mathematical framework to render the control problem formulation numerically tractable for practical applications is summarized based on different contributions that have appeared in literature over the years. In order to shield practicing control engineers from these low-level mathematics and to assist them throughout the complete typical design procedure of a linear control system, we introduce a software package that is demonstrated by example.

On the other hand, convincing results on four mechatronic case studies are presented to demonstrate the systematic work flow up to the level where automated software tools can handle the remainder of the work. These include an induction motor of a back-to-back test rig, an educational setup with a flexible link, a rotordynamic test rig with 5 active magnetic bearings, and an industrial AGV. In every case study, the attention is drawn to specific benefits of modern linear feedback control tools as compared to the classic approaches that are widely accepted. Indeed, besides a systematic design flow, closed-loop controller synthesis allows for parametric optimization, takes full advantage of the model and can directly account for robustness requirements. Such properties are valuable for many real-world control problems, certainly in mechatronics.

An analysis of the strengths, weaknesses, opportunities and threats ('SWOT') of the proposed design methodology gives an overview on what has been achieved and what is suggested as future work to further improve the acceptance of modern linear feedback control in industry.