Nonlinear Attitude Filter on SO(3): Fast Adaptation and Robustness

TL;DR

This paper introduces a fuzzy-rule-based adaptive nonlinear attitude filter on SO(3) that achieves fast convergence and robustness by online tuning of the adaptation gain using artificial bee colony optimization.

Contribution

It presents a novel functional approach for online adaptation gain tuning in nonlinear attitude filters using fuzzy rules and optimization, enhancing convergence and robustness.

Findings

High adaptation gain at large errors improves convergence.

Small adaptation gain at small errors enhances robustness.

Simulation shows robustness against large initial errors and measurement uncertainties.

Abstract

Nonlinear attitude filters have been recognized to have simpler structure and better tracking performance when compared with Gaussian attitude filters and other methods of attitude determination. A key element of nonlinear attitude filter design is the selection of error criteria. The conventional design of nonlinear attitude filters has a trade-off between fast adaptation and robustness. In this work, a new functional approach based on fuzzy rules for on-line continuous tuning of the nonlinear attitude filter adaptation gain is proposed. The input and output membership functions are optimally tuned using artificial bee colony optimization algorithm taking into account both attitude error and rate of change of attitude error. The proposed approach results of high adaptation gain at large error and small adaptation gain at small error. Thereby, the proposed approach allows fast convergence properties with high measures of robustness. The simulation results demonstrate that the proposed approach offers robust and high convergence capabilities against large error in initialization and uncertain measurements.