Distributed Absolute and Relative Estimation of Spacecraft Formations

Abstract

For optimal decision making and collaborative task performance in multispacecraft missions, it is essential for each spacecraft to maintain a state estimate of all the spacecraft in the formation. However, especially in the case of large formation sizes, each spacecraft may not be able to track or communicate with all other spacecraft. Further, for formations deployed in deep space, the unavailability of the Global Navigation Satellite System makes inertial state estimation challenging. We propose the Distributed Absolute and Relative Estimation (DARE) algorithm for autonomous inertial estimation of spacecraft formations. The algorithm enables each spacecraft to maintain an accurate inertial estimate of the entire formation even in the presence of observability and communication constraints. Each spacecraft utilizes measurements of feature points (landmarks) for inertial localization and measurements of neighboring spacecraft for relative localization. The algorithm is distributed and scalable to any number of spacecraft in the formation. A modified version of the algorithm called the Sparse Distributed Absolute and Relative Estimation (SDARE) algorithm is also derived. This algorithm is computationally more efficient at the expense of estimation accuracy, making it suitable for implementation on nanosatellites where resources are limited. Numerical simulation results demonstrating the effectiveness and comparing the performance of both algorithms are provided.