Attitude Data Fusion Using a Modified Rodrigues Parametrization

Data fusion in the presence of unknown correlations is an expanding area of research as many space related applications move toward smaller, distributed systems. While nontrivial by nature, the problem is further compounded when one considers fusion of attitude estimates where singularities and constraints must be accounted for. Prior work in this area has considered fusion when the attitude is parameterized using the four-dimensional quaternion. While the quaternion has many advantages, its norm constraint provides some added difficulty to the fusion process which may render it unsuitable for onboard spacecraft attitude fusion systems. This work provides an alternative means for data fusion by parameterizing the attitude using the three-dimensional modified Rodrigues parameters. Because they represent a minimally parameterized set, modified Rodrigues parameters fall victim to a singularity which must be considered during fusion. Two algorithms, a global and a local approach are developed within the Covariance Intersection framework for fusing modified Rodrigues parameters while avoiding the singularity. Simulation results validate the approach and show a computational advantage over previous quaternion based algorithms.