Sequential Monte Carlo Methods for Spacecraft Attitude and Angular Rate Estimation

אבישי כרמי

Department of Mechanical Engineering
Ariel University Center And Asher Space Research Institute, Technion

Sequential Monte Carlo (SMC) methods, otherwise known as particle filters, are a family of simulation-based nonlinear filtering techniques that have drawn much attention in the last decade. By virtue of their recursive Bayesian structure and unique sampling mechanization, these methods have proved themselves highly viable and accurate in various signal processing applications. In this talk we present a novel class of SMC algorithms for spacecraft attitude and angular rate estimation from noisy vector observations. The class includes: (1) a novel quaternion particle filter (QPF), (2) an adaptive variant of the QPF, which is capable of recovering either inaccurate or unknown sensor noise statistics, (3) a fast and robust attitude-free angular rate particle filter for gyroless spacecraft applications, and (4) a novel attitude and angular rate particle filter. An extensive comparison of the new filtering methods with state-of-the-art attitude estimation techniques demonstrates their decisive superiority in terms of accuracy, robustness, and rate of convergence. The remaining part of this talk is devoted to some theoretical aspects pertaining to the attitude estimation problem itself. As part of this we shed light on the nature of the quaternion estimation error covariance, a controversial subject that has ignited a long running debate back in the 80's.

ההרצאה תתקיי ביום רביעי י"א בתמוז תש"ע‬ (23.06.10)
שעה‬ 16:30
בנין אוירונוטיקה חדר‬ 241
כבוד קל יוגש לפני ההרצאה‬