High Performance Control and Extended Kalman Filter Based Estimation of Sensorless Permanent Magnet Synchronous Motor Drive for Robotic Applications

Abstract

Owing to high efficiency and high performance controllability, permanent magnet synchronous motors (PMSM) are being considered for various robotic applications including the Articulated Robots. PMSM drive includes a position/speed sensor for self synchronous control. Sensorless operation refers to the possibility of removing the afore-mentioned position/speed sensor to increase the robustness of the system and improve the reliability against sensor failures. This requires speed/position estimation for closed loop control. Extended Kalman filter (EKF) is a viable option as an observer owing to its noise rejection characteristics, ease in tuning the observer and the recursive algebraic nature of the algorithm which translates to real time hardware implementability. This paper proposes sensorless PMSM drive for robotic applications, with the dual perspective of improving the control and estimation aspects of the drive. A modification to the speed controller of Proportional+Integral (PI controller) type is proposed, wherein the overshoot in speed is drastically reduced without the necessity of including the Differential control. Likewise, a Proportional+Differential (PD) type is proposed for position control, with a suggestion to reduce the peak overshoot in position response. An adaptive Kalman filter is proposed to improve the estimation of speed/position, to achieve a high performance closed loop controlled PMSM drive.