Analysis of Complexity Reduction in Kalman Filters Through Decoupling Control With Chattered Inputs in PMSM

Research output: Contributions to collected editions/worksArticle in conference proceedingsResearchpeer-review


The purpose of this paper is to investigate the idea that by implementing decoupling control algorithms in Permanent Magnet Synchronous Machine (PMSM) controllers, observers like extended Kalman filters (EKFs) can achieve an advantage in computational load. In particular, the KF is designed as a combined state and parameter estimator. The approach is based on the possibility to consider individual electrical current dynamics, together with key system parameters influencing their dynamics, i.e. their inductances Ld, Lq, which is enabled by a decoupling control. Even though the decoupled dynamics are linear, the resulting augmented subsystems, including the inductances as estimation variables, are nonlinear. Furthermore, the paper shows that high-frequency inputs, caused, for example, by control chattering, can enable the EKF-based estimation of the inductances by providing for a never-ending transient phase in the case of the decoupled variant. The computational advantage is ultimately achieved by reducing the complexity of the Kalman filters by around 75%. The results are validated using computer simulations of coupled as well as decoupled control systems, also demonstrating the benefits of chattery input voltages.
Original languageEnglish
Title of host publication15th European Workshop on Advanced Control and Diagnosis (ACD 2019) : Proceedings of the Workshop Held in Bologna, Italy, on November 21–22, 2019
EditorsElena Zattoni, Silvio Simani , Giuseppe Conte
Number of pages13
Place of PublicationCham, Switzerland
PublisherSpringer Nature Switzerland AG
Publication date2022
ISBN (Print)978-3-030-85317-4
ISBN (Electronic)978-3-030-85318-1
Publication statusPublished - 2022
EventConference - 15th European Workshop on Advanced Control and Diagnosis (ACD 2019) - Alma Mater Studiorum, Bologna, Italy
Duration: 21.11.201922.11.2019
Conference number: 15