This paper presents a robustness analysis of three state estimation methods for sensorless control of permanent magnet synchronous motors (PMSM): a conventional back electromotive force (BEMF) observer, a gradient-based adaptive flux observer, and a nonlinear observer based on Dynamic Regressor Extension and Mixing (DREM). Their theoretical foundations and convergence properties are analyzed. Simulation and experimental tests under measurement noise, parameter perturbations, and load disturbances evaluate steady-state error, dynamic response, and robustness. Results show that the conventional observer is parameter-sensitive, the adaptive observer achieves asymptotic convergence, while the DREM-based observer provides faster convergence.

Сюй Ч. (науч. рук. Ведяков А.А.) Robustness Investigation of State Estimation Methods for PMSM // Сборник тезисов докладов конгресса молодых ученых. Электронное издание. – СПб: Университет ИТМО, [2026]. URL: https://kmu.itmo.ru/digests/article/17290