AI-Based Virtual Sensor for Rotor Position Estimation

Learn how to estimate the rotor position of a permanent magnet synchronous motor (PMSM) using an AI-based virtual sensor, eliminating the need for physical speed/position sensors. The approach leverages deep learning to generate, train, test, and deploy a neural network that predicts the rotor’s electrical position from voltage and current measurements.

Walk through the complete workflow using a Motor Control Blockset™ reference example:

• Generate training data through field-oriented control simulation.
• Create and train the neural network with Deep Learning Toolbox™.
• Validate the trained network using simulation.
• Generate code and deploy the trained virtual sensor to target hardware.

This demonstration is for motor control engineers seeking AI-based alternatives to traditional sensors, offering a scalable and cost-effective approach to rotor position estimation.

See example: Field-Oriented Control of PMSM Using Position Estimated by Neural Network