PhD Aeronautics & Astronautics, Stanford University, In Progress
MS Aeronautics & Astronautics, Stanford University, 2012
BS Aeronautics & Astronautics, University of Washington, 2010
Asteroids are vital to our understanding of the evolution of our solar system, but current ground-based techniques are unable to accurately characterize their size, shape, composition, and local geology. In-situ measurements (aka, sending a spacecraft!) are necessary.
My research focuses on techniques for creating accurate shape models of asteroids, using a LiDAR as a sensor. Previous asteroid missions have depended on cameras as their main mapping sensor, which can leave large areas of the asteroid unmappable if they are in shadow for the duration of the mission. Using LiDAR as the main mapping sensor alleviates this issue, but poses another problem: if the asteroid has drifted substantially between multiple measurements of the same region, how can the LiDAR's point cloud measurements be properly realigned again?
My work draws upon computer vision techniques, star tracker algorithms, and graph matching algorithms to autonomously identify similar regions of a point cloud. These regions can then be properly realigned using GraphSLAM or some other appropriate algorithm. This will not only be useful for future spacecraft missions, but has been demonstrated on an underwater seafloor mapping expedition, and is intended for use on an iceberg expedition in 2017.
- Spacecraft Guidance, Navigation, & Control
- Computer Vision
- Image Processing
- Machine Learning
Last modified Thu, 29 Sep, 2016 at 19:26