Error-driven Tracking Control Design with Preset-input LMS Adaptive Filter

There is a problem of insufficient reference tracking accuracy in servo system driven by permanent magnet synchronous motor (PMSM), which limits its application. This paper aims to propose a modified error-driven preset-input least mean square adaptive filter (EPLMS-AF) to improve the tracking accuracy. The reference tracking error is used to drive the update of adaptive weights in EPLMS-AF. The designed preset-inputs makes the reference tracking error converge to zero according to the form of reference. On the premise of choosing an appropriate iteration size to ensure the stability of the filter, the tracking error is able to reduce to zero along the gradient direction. At last, numerical simulations are provided to illustrate the high efficiency and wide applicability of the proposed method.

1 Introduction

Permanent magnet AC servo system control technology plays an important role in aerospace, industrial robot and other fields [1]–[3]. Permanent magnet synchronous motor (PMSM) is widely used as the actuator of servo system since it has fine control performance and high efficiency. For servo system with PMSM, one of its important properties is control accuracy, including angular velocity stability and angular position accuracy [4]. However, the servo system always has reference tracking error if the angular velocity reference is periodic or unidirectional. The reference tracking error in the servo process is difficult to eliminate from the perspective of design process or mechanical manufacturing [5]. This limits the application of PMSM. Therefore, it is necessary to improve the reference tracking accuracy by control methods.