A Monte-Carlo performance analysis of Kalman filter and targeting algorithms for autonomous orbital rendezvous

Abstract

Autonomous orbital rendezvous with an orbiting sample (OS) is seen as an enabling technology for a Mars Sample Return (MSR) mission, so several demonstrations have been planned. With CNES cooperation a proposed rendezvous demonstration was governed by ITAR restrictions, and a guidance and navigation system was designed using a Precomputed Gain Kalman filter and targeting algorithms. Having lost CNES participation, the opportunity now exists to use a full Extended Kalman filter with onboard targeting algorithms on a new demonstration using the Mars Telecommunications Orbiter (MTO). This creates an impetus to compare the Precomputed Gain system with the Extended system to determine their relative performance. This thesis aims to compare the Precomputed Gain and Extended Kalman filters and associated targeting algorithms using a Monte-Carlo analysis, and based on quantitative performance metrics including: total change in velocity required, navigation errors, target pointing errors. In addition, other aspects of the algorithms will be studied including: technology readiness level (TRL) data uplink requirements, and complexity and computational burden for the onboard algorithms. Monte-Carlo analysis will reveal that the Extended system modestly outperforms the Precomputed Gain system in total change in velocity required, navigation error, and target pointing error, with a larger performance envelope. The Extended system will also be found to have a greater technology readiness and require substantially less data uplink. The Precomputed Gain system will be found to be a significantly less complex algorithm for the onboard flight computer.