An advanced sensorless control strategy using PS-EKF-based Integral Backstepping Controller for PMSM : Experimental validation using dSPACE DS1104

High-performance sensorless (SL) control of Permanent Magnet Synchronous Motors (PMSMs) remains a crucial challenge in modern industrial applications, particularly in harsh environments where mechanical sensors are impractical. In this paper, we propose an advanced SL position and speed (PS) approach combining an extended Kalman filter (EKF) with an Integral Backstepping controller (IBSC). The developed SLPS-IBSC-EKF method eliminates the need for position and speed sensors while maintaining high dynamic performance. To achieve this, the EKF with its optimal statistical formulation and adaptability to non-linear systems ensures precise reconstruction of rotor PS even in the presence of noise. Meanwhile, the IBSC, through its recursive design based on Lyapunov stability criteria, adaptively compensate for PMSM non-linearities, guarantees stability against external disturbances, and addresses parametric uncertainties. Furthermore, incorporating integral action significantly reduces steady-state tracking errors, improves disturbance rejection, and enhances overall control performance. Comparative analysis in Matlab/Simulink shows that the proposed method surpasses the SLPS-PI-EKF technique across several key performance metrics, including Settling time, Rise time, Rejection time, Steady-State Error and Torque ripple (peak-to-peak). These improvements represent enhancements of 91.7 %, 90.9 %, 89.9 %, 85.6 %, and 63.4 % respectively, achieved by the SLPS-IBSC-EKF compared to the SLPS-PI-EKF. Furthermore, a performance comparison table with other recent methods from the literature reinforces the advantages of the proposed method. Finally, the SLPS-IBSC-EKF is implemented and tested on a small-scale physical test bench using the DSPACE DS1104 board. The results from these tests confirm excellent correlation with simulations and demonstrate the practical viability of the proposed method for real-time implementation.