Estimation within the CLARAty Architecture
- Summary
The goal of this effort is to support the design, implementation and documentation of the estimation package within the CLARAty architecture. This infrastructure package will provide a uniform solution to allow various types of estimation algorithms to be implemented across a wide range of robotic vehicles, e.g. Rocky 7 (JPL), Rocky 8 (JPL), FIDO (JPL), K9 rovers (NASA Ames). We derive different layers of abstractions from a wide range of current implementations of estimators in the field of robotics, primarily for localization and mapping, and provide classifications for essential estimation components that will lead to generic implementations of them with reduced variability. A significant impact of this work is that it will allow for code portability and reusability and therefore reduce development cost by sharing the same functionality on more than one vehicle. In addition, it will provide the necessary software architecture to support competing estimators and allow for direct comparison of their performance on the same robot platform. This effort is realized in close collaboration with leading robotics centers at the Jet Propulsion Laboratory, the NASA Ames Center, Carnegie Mellon University, University of Michigan, and the Massachusetts Institute of Technology.
