My name is Jacob Cassady. I work at NASA Langley Research Center in the Flight Software Systems Branch. My current focus is autonomy; most of my time is spent as co-lead for the software and autonomy of the Tall Lunar Tower mission.

• Machine Learning
• Robotics
• Computer Vision
• Control Systems

The increased use of image classification in technology today has incited a rise in research and development for new approaches in facial detection and identification models. Two common problems in image classification are storing large datasets and model training costs. One approach to achieving dimensionality reduction while maintaining performance is Principal Component Analysis where a subset of eigenvectors, also known in the domain of facial detection as ``eigenfaces'', are used to represent the data in a lower dimensionality space. This paper presents an image classification model based on eigenfaces and support vector machines using the Amsterdam Dynamic Facial Expression Set (ADFES) dataset. Implementation of an image classification model is described, and performance analysis of the model is presented with a focus on the efficacy of using eigenfaces when training the model.

Notes on the first 4 chapters of Neural Network Design by Hagan, et al. I don't plan to finish but a lot of people have found value in them.

My first big project from my first programming class. Sometimes it's helpful to see how rough everyone's start is. It might look ugly; but I bet you cant beat it : )

TAALE is a compact, lightweight, and portable, low-power self-erecting tower for use on landers, exploration rovers, and robotic lunar surface operations. Because the lunar tower is multi-functional and autonomous, it can help close two capability gaps: over horizon communications; and landing within 50 meters of a specified landing site on the Moon. This project aims to demonstrate autonomous erection of a fixed lunar infrastructure, develop systems to improve the lunar towers to support payloads, demonstrate the creation of a local WiFi network for data transfer from the top of the tower to the bottom, and demonstrate a stable platform with power and data routing for landing site survey payloads. [Taken from https://www.nasa.gov/feature/nasa-empowers-workforce-to-advance-deep-space-technologies]

• Role: Software and Autonomy Lead
• Mission Directorate: Research Directorate

Stewart platform based modular manipulator. See this public description: https://www.nasa.gov/feature/langley/nasa-assemblers-are-putting-the-pieces-together-for-autonomous-in-space-assembly

• Role: Software Engineer, autonomy team
• Mission Directorate: Research Directorate

NASA’s Low-Earth Orbit Flight Test of an Inflatable Decelerator, or LOFTID, is demonstrating a cross-cutting aeroshell -- a type of heat shield -- for atmospheric re-entry. For destinations with an atmosphere, one of the challenges NASA faces is how to deliver heavy payloads (experiments, equipment, and people) because current rigid aeroshells are constrained by a rocket’s shroud size. One answer is an inflatable aeroshell that can be deployed to a scale much larger than the shroud. This technology enables a variety of proposed NASA missions to destinations such as Mars, Venus, Titan as well as return to Earth. [Taken from https://www.nasa.gov/loftid]

• Role: Software Engineer, ground system
• Mission Directorate: Space Technology Mission Directorate

One of the most challenging tasks in remote sensing from space is achieving required instrument calibration accuracy on-orbit. The Moon is considered to be an excellent exoatmospheric calibration source. However, the current accuracy of the Moon as an absolute reference is limited to 5 – 10%, and this level of accuracy is inadequate to meet the challenging objective of Earth Science observations. ARCSTONE is a mission concept that provides a solution to this challenge. An orbiting spectrometer flying on a small satellite in low Earth orbit will provide lunar spectral reflectance with accuracy sufficient to establish an SI-traceable absolute lunar calibration standard for past, current, and future Earth weather and climate sensors. [Taken from https://science.larc.nasa.gov/ARCSTONE/]

• Role: Software Engineer, developed controller and simulator for cryocooler
• Mission Directorate: Science Directorate

• Jacob T. Cassady, Matthew Mahlin, Emma Kravets, Matthew Vaughan and Matthew Rodgers. "Software Design for the Supervised Autonomous Assembly of a Tall Lunar Tower," AIAA 2023-4682. ASCEND 2023. October 2023.
• N. H. James, J. R. Cooper, M. K. Mahlin, L. White, J. Cassady, J. Mulvaney, M. P. Vaughan and I. Wong. Assemblers Project Review: Building and Testing an Autonomous Modular and Reconfigurable Manipulation System. In 2022 AAIA SciTech Forum, January 3 – 7, 2022, San Diego, CA, USA.
• Kyongchan Song, Martin Mikulas, Matthew K. Mahlin, Jacob T. Cassady. Sizing and Design Tool for Tall Lunar Tower. in 2023 AAIA SciTech Forum, January 23 – 27, 2023, National Harbor, MD, USA.
• Jacob T. Cassady, Chris Robinson, Dan O. Popa. 2020. Increasing user trust in a mobile robot using explainable AI in a traded control paradigm. In The 13th PErvasive Technologies Related to Assistive Environments Conference (PETRA ’19), June 30 – July 3, 2020, Corfu Island, Greece. ACM, New York, NY, USA.

• M.S. Robotics and Autonomous Systems, Johns Hopkins Whiting School of Engineering [IN PROGRESS]
• B.S. Computer Science and Engineering, honors, University of Louisville J.B. Speed School
• B.S. Electrical Engineering, honors, University of Louisville J.B. Speed School

• English
• Español
• Français