Integral Compensation Function Observer and Its Application to Disturbance-Rejection Control of QUAV Attitude

Observer-based disturbance rejection holds substantial theoretical and practical relevance in the field of control engineering, with numerous variants of disturbance observers already schemed. Nevertheless, the criteria for accuracy and avenues for enhancement remain areas warranting further investigation. This article introduces an integral compensation function observer (CFO) featuring a novel structure and efficient utilization of information for estimating disturbances in
$n$
th-order uncertain systems. This approach enhances estimation accuracy by addressing the inherent limitations of the linear extended state observer (LESO), such as low order, lacking usage of information, nonconvergence, and limited bandwidth. Through the derivation and quantification of the disturbance sensitivity transfer function (DSTF), this study examines the disturbance sensitivities of the CFO, LESO, and an improved ESO (IESO). The findings indicate that the CFO elevates the estimable order of disturbance and surpasses both LESO and IESO in bandwidth and disturbance estimation accuracy. In evaluating both the estimation accuracy of disturbance of the CFO and the disturbance-rejection performance (DRP) of CFO-based control, nonlinear pole assignment controls (NPACs) employing 2nd/3rd-order CFO, IESO, LESO, and 4th-order CFO are implemented in the context of attitude control for a quadrotor unmanned aerial vehicle (QUAV) that is exposed to prearranged disturbance torques. The results illustrate that the CFO outperforms the IESO and LESO in terms of accurately estimating the prearranged disturbing torques. Furthermore, the recorded magnitudes of attitude in response to disturbances underscore the superior DRP of CFO-NPAC relative to IESO-NPAC and LESO-NPAC.