Dual quaternion kalman filters for spacecraft relative pose estimation

This work presents two kinematic additive dual quaternion Kalman filters for position, attitude and sensor biases estimation in a low Earth orbit two-spacecraft relative navigation problem. The spacecraft are assumed to be equipped with sensing suites that are suitable for close proximity operations (within the meters scale), which provide biased angular and linear velocities, and a batch of inter-satellite line-of-sight observations extracted from a vision-based navigation system. Three other different algorithms, additive and multiplicative, are developed involving dual quaternions and standard pose representations. Extensive numerical simulation is performed for a proximity operation scenario and the filters' performances are compared.