Abstract: This paper delves into sensorless control techniques for AC motors, critical for reducing system complexity and improving reliability. We review observer-based methods (sliding mode, Luenberger observers) and artificial intelligence (AI)-driven approaches using neural networks. Challenges such as parameter sensitivity, low-speed operation, and noise immunity are addressed through hybrid control architectures. Real-time implementation on FPGAs and digital signal processors (DSPs) is discussed, including field-oriented control (FOC) without position sensors.
Content Highlights:
High-Frequency Injection Methods: Analysis of rotor saliency tracking for interior PMSMs (IPMSMs).
AI-Enhanced Observers: Training recurrent neural networks (RNNs) for online parameter estimation.
Hardware-in-the-Loop (HIL) Testing: Validation of sensorless algorithms under dynamic load conditions.
Comparison of Sensorless Techniques: Performance metrics for induction motors vs. PMSMs in automotive applications.
