A Novel Data-Driven Attitude Estimation: Retrospective Cost Attitude Filtering (RCAF)

Abstract

Attitude filtering is a critical technology with applications in diverse domains such as aerospace engineering, robotics, computer vision, and augmented reality. Although attitude filtering is a particular case of the state estimation problem, attitude filtering is uniquely challenging due to the special geometric structure of the attitude parameterization. This thesis presents a novel data-driven attitude filter, called the retrospective cost attitude filter (RCAF), for the SO(3) attitude representation. RCAF uses a multiplicative correction signal computed using retrospective cost optimization and measured data. The RCAF filter is numerically and experimentally validated in a scenario with noisy attitude measurements and noisy and biased rate-gyro measurements.