1: \begin{abstract}

2: Nonlinear attitude filters have been recognized to have simpler structure

3: and better tracking performance when compared with Gaussian attitude

4: filters and other methods of attitude determination. A key element

5: of nonlinear attitude filter design is the selection of error criteria.

6: The conventional design of nonlinear attitude filters has a trade-off

7: between fast adaptation and robustness. In this work, a new functional

8: approach based on fuzzy rules for on-line continuous tuning of the

9: nonlinear attitude filter adaptation gain is proposed. The input and

10: output membership functions are optimally tuned using artificial bee

11: colony optimization algorithm taking into account both attitude error

12: and rate of change of attitude error. The proposed approach results

13: of high adaptation gain at large error and small adaptation gain at

14: small error. Thereby, the proposed approach allows fast convergence

15: properties with high measures of robustness. The simulation results

16: demonstrate that the proposed approach offers robust and high convergence

17: capabilities against large error in initialization and uncertain measurements. 

18: \end{abstract}