For the 71st IAC: Implementation and Validation of Murrell’s Version Kalman Filter for Attitude Estimation

Abstract

Cubesats with imaging payloads face unique challenges in terms of stringent pointing accuracy and stability requirements. Team Anant is a student-run technical team working to build a 3U Cubesat. This paper discusses the implementation, validation and integration of an attitude estimation algorithm as part of the satellite’s Attitude Determination System (ADS). The ADS hardware usually comprises sensors such as an IMU, magnetometer, and sun sensors. Validation methodology and architecture design, which aims to satisfy the allocated pointing budget, are also discussed. The paper introduces the motivation to choose Murrell’s version Kalman Filter and a comparison with popular alternatives. This is followed by some prerequisites, after which, the paper describes the top level overview and testing framework developed for the Kalman Filter. This requires emulating the in-orbit environment and tracking the true state to establish the performance limit with a predefined performance metric. The verification procedure adopted by the team is discussed in detail. Apart from analysing the expected trend of the filter parameters over time, a quasi-Monte Carlo approach was also followed. Furthermore, the Cramer–Rao bound is used to establish a lower bound on the error covariance matrix. Lastly, an approach for fine sensor selection is provided based on emulating its integration with the ADS. The paper concludes by discussing the lessons learnt and the important stages in the development and testing of an attitude estimation algorithm.