Attitude estimation algorithm based on rodrigues parameter

A fast all-attitude estimation algorithm is proposed using Rodrigues parameter to describe the attitude. Rodrigues parameter is simple and efficient in the problem of attitude estimation, but it cannot be used in large-angle motions, because it will be singular. A simple and non-singularity attitude representation is presented by combining the method of sequential rotations and Rodrigues parameter. Then, the problem of attitude estimation using the presented attitude representation for the typical attitude determination mode composed of vector observations and gyros is discussed. In the process of designing attitude estimation algorithm, the gyros model and vector observation model are provided in the first step. Subsequently the state and covariance propagation and update equations are derived. Then an efficient integration method for state and covariance propagation is presented. Simulation results demonstrate that the attitude estimation accuracy using the new method is equivalent to that using the quaternion estimation algorithm, but the former shows a reduction of about 10% in computational load.