Andrew Spielvogel presented his thesis research entitled “Online Inertial Measurement Unit Sensor Bias and Attitude Estimation For The Calibration And Improved Performance Of Attitude And Heading Reference Systems” on March 6, 2020. His thesis committee included Professor Louis Whitcomb (primary advisor), Assistant Professor Marin Kobilarov, and Professor Noah Cowan.
Dynamic instrumentation and estimation of vehicle attitude is critical to the accurate navigation of land, sea, and air vehicles in dynamic motion. The focus of this thesis is the development of algorithms for improved performance of attitude and heading reference systems (AHRSs) and robotic vehicle navigation. Inertial measurement unit (IMU) sensor bias estimation methods for use in the calibration of AHRS and an adaptive attitude estimator operating directly of $SO(3)$ are reported. The reported algorithms provide online calibration and attitude estimation methods which enable more accurate navigation for robotic vehicles.
This thesis differentiates AHRSs into two categories: AHRSs that estimate true-North heading and those that estimate magnetic north heading. First, this thesis report several novel algorithms for micro-electro-mechanical systems (MEMS) IMU sensor bias estimation. Observability, stability, and parameter convergence are evaluated in numerical simulations, full-scale vehicle laboratory experiments, and full-scale field trials in the Chesapeake Bay, MD. Second, this thesis reports an adaptive sensor bias observer and attitude observer operating directly on SO(3) for true-North gyrocompass systems that utilize six-degree-of-freedom IMUs with three-axis accelerometers and three-axis angular rate gyroscopes (without magnetometers) to dynamically estimate the instrument’s time-varying true-North attitude (roll, pitch, and geodetic heading) in real-time while the instrument is subject to a priori unknown rotations. Stability proofs for the reported bias and attitude observers, preliminary simulations, and a full-scale vehicle trial are reported.
The reported calibration and attitude estimation methods are shown experimentally to improve calibration of AHRS attitude estimation over current state of the art sensor bias estimation methods, and this thesis presents a true-North gyrocompass system based on adaptive observers for use with strap-down IMU. These results may prove to be useful in the development of navigation systems for small low-cost robotic vehicles.
Andrew R. Spielvogel was born in Boston, Massachusetts in 1990. From 2004-2009, he attended the Green Mountain Valley School where he was Valedictorian of the Class of 2009. Afterwards, he attended Harvard University, where he graduated cum laude in May 2013 with a S.B. in Electrical Engineering with high honors in field. While at Harvard, Andrew competed for four years on the NCAA Division 1 Harvard Alpine Ski team where he was named three times to the All-Academic Team (2011-2013) and twice captain of the ski team (2011-2012, 2012-2013). After graduating college, Andrew was an Electrical Engineering Intern at Harvest Automation where he worked on a vision system for mobile robots in the nursery and greenhouse industries. In August 2014, he enrolled in the Mechanical Engineering Ph.D. program at Johns Hopkins University, where he received a Robotics MSE in 2017.