Semester

Fall

Date of Graduation

2021

Document Type

Dissertation

Degree Type

PhD

College

Statler College of Engineering and Mineral Resources

Department

Mechanical and Aerospace Engineering

Committee Chair

Jason N. Gross

Committee Member

Yu Gu

Committee Member

Guilherme A. S. Pereira

Committee Member

Natalia A. Schmid

Committee Member

Powsiri Klinkhachorn

Committee Member

Kyohei Otsu

Abstract

Accurate localization is a critical component of any robotic system. During planetary missions, these systems are often limited by energy sources and slow spacecraft computers. Using proprioceptive localization (e.g., using an inertial measurement unit and wheel encoders) without external aiding is insufficient for accurate localization. This is mainly due to the integrated and unbounded errors of the inertial navigation solutions and the drifted position information from wheel encoders caused by wheel slippage. For this reason, planetary rovers often utilize exteroceptive (e.g., vision-based) sensors. On the one hand, localization with proprioceptive sensors is straightforward, computationally efficient, and continuous. On the other hand, using exteroceptive sensors for localization slows rover driving speed, reduces rover traversal rate, and these sensors are sensitive to the terrain features. Given the advantages and disadvantages of both methods, this thesis focuses on two objectives. First, improving the proprioceptive localization performance without significant changes to the rover operations. Second, enabling adaptive traversability rate based on the wheel-terrain interactions while keeping the localization reliable.

To achieve the first objective, we utilized the zero-velocity, zero-angular rate updates, and non-holonomicity of a rover to improve rover localization performance even with the limited available sensor usage in a computationally efficient way. Pseudo-measurements generated from proprioceptive sensors when the rover is stationary conditions and the non-holonomic constraints while traversing can be utilized to improve the localization performance without any significant changes to the rover operations. Through this work, it is observed that a substantial improvement in localization performance, without the aid of additional exteroceptive sensor information.

To achieve the second objective, the relationship between the estimation of localization uncertainty and wheel-terrain interactions through slip-ratio was investigated. This relationship was exposed with a Gaussian process with time series implementation by using the slippage estimation while the rover is moving. Then, it is predicted when to change from moving to stationary conditions by mapping the predicted slippage into localization uncertainty prediction. Instead of a periodic stopping framework, the method introduced in this work is a slip-aware localization method that enables the rover to stop more frequently in high-slip terrains whereas stops rover less frequently for low-slip terrains while keeping the proprioceptive localization reliable.

Download

Share

COinS