The pre-recorded talks and posters on this page showcase the work of students who received NC Space Grant research funding for the 2019-20 and 2020-21 academic years. The menu at right provides links to pre-recorded talks and posters by other funded students on additional topics.

Ryan Daugherty 

NASA Internship Award at Goddard Space Flight Center – Summer 2020
North Carolina State University
Undergraduate Student (Senior), Computer Engineering

Bundle Protocol Node Dependency Repository 

The goal of my project was to simplify the installation process of the Bundle Protocol Node Ground (BPN-GND) package. The installation of BPN-GND could take days for a NASA engineer, wasting valuable time when onboarding a new hire or setting up a new machine. The complex installation process is compounded by the numerous dependencies that BPN-GND requires, many of which have unclear documentation. My project was to figure out all the dependencies BPN-GND requires, install them all into a local directory, then save that directory in version control software (in this case, Git). As I installed each dependency, I also wrote documentation detailing how I installed each dependency in case the dependency needs to be updated or replaced; the completed documentation was stored in the repository with the dependencies themselves. The finished repository was tested by my mentor to ensure all the dependencies were installed correctly and functioned with BPN-GND. Without the repository, a NASA engineer could spend a few days navigating the tangled web of dependencies and open source programs needed to run BPN-GND. With the repository, that process is reduced to the few minutes it takes to clone the repository from Git. The Bundle Protocol Node Dependency Repository represents a massive quality of life improvement for the engineers working on Bundle Protocol Node Ground.

Advisor: Robert Wiegand (NASA Software Developer)

Arjon Zeus Del Rosario 

NASA Internship Award at Langley Research Center – Spring 2021
The University of North Carolina Wilmington
Recent Graduate (December 2021), Computer Science

Enabling Laser Powder Bed Fusion (Software Development) (Pre-recorded talk)

I am a NASA Spring Intern working on software development for enabling laser powder bed fusion(LPBF). I am not allowed to talk about the project in detail yet, until it has been approved for release. But, the presentation that I will most likely give is about the process of rendering 3D objects using 3D printers and how it is done efficiently and effectively using software tools.

Advisor: Samuel Hocker, NASA Langley Research Center

Joshua Hilbish

2020-2021 NC Space Grant Undergraduate Research Scholar
Fayetteville State University
Undergraduate Student (Senior), Computer Science

Detection of Defects in Manufacturing Processes 

This project has developed machine learning models that can be used to automate the detection of defects in manufacturing processes and to predict machining operating regimes which can produce manufacturing defects. An image analysis neural network based on Mask R-CNN was developed to perform instance segmentations on surfaces images of artifacts in free form optics. The microscopic images of these milled lenses may show tooling blemishes of different shapes and clustering patterns. The detection of these defects through image segmentation by the neural network can help characterize the quality of the manufactured lenses. In this project manually annotated images were used to train the neural network and its performance was measured using held-out annotated images. The current results of the project make it clear that to increase the performance of the neural network larger data sets will need to be collected, especially with the future goal of deploying the model on edge compute devices. The goal of the predictive neural network modeling is to analyze sound data from milling to classify the manufacturing process as stable or self-vibration (known as chatter). This predictive model can help reduce the waste of material in manufacturing by automatically selecting machining parameters that result in stable milling. Software packages were used to convert raw milling recordings to waveforms, and to compute spectrograms which can be processed with a convolutional neural network. Experiments are underway to compare the performance of the neural network when trained on the raw data with a different version of the neural network trained on the spectrogram features.

Faculty Advisor: Sambit Bhattacharya

Finn James 

2020-2021 NC Space Grant Undergraduate Research Scholar
The University of North Carolina Chapel Hill
Undergraduate Student (Junior), Computer Science

Using Automated Robust Chauvenet Rejection to Reduce Noise in Radio Spectrum Binary Black Hole System Images 

The OJ 287 binary black hole system’s periodic brightening and dimming pattern presents a unique opportunity to test gravitational wave models of interacting black holes; however, the last brightening event occurred while the system was very near to the sun, so it was difficult for anything but radio frequency imaging to observe it. Although this data could hold new insights into General Relativity, its noisiness and strong solar interference prohibit immediate interpretation of the dimming of the system. My team and I have created a new automated system for running Robust Chauvenet Rejection (RCR) on noisy astronomical data. RCR applies cascading outlier rejection methods, transitioning from most robust and least precise to most precise and least robust, so that the algorithm is both robust and precise even with data sets dominated by noise. By creating this automated system, RCR can be integrated with the Skynet robotic telescope network, thus allowing scientists and students globally to use RCR to process noisy images. Furthermore, this new system has been successful in dramatically reducing noise in the OJ 287 dataset.

Faculty Advisor: Dan Reichart

Cade Justad-Sandberg

NASA DEVELOP Internship Award – Fall 2019
East Carolina University
Graduate Student (Masters), Computer Science

Assessing Vegetation Response to Remote Sensing Drought Indices within the Dry Corridor of Central America using NASA Earth Observations 

The dry corridor resides primarily in the pacific region of Central America, which experiences severe drought during the El Niño Southern Oscillation cycle. El Niño causes severe climate variances in Central America that impact agriculture, livelihoods, and hydrological cycles. The region is rich with cash crops, including bananas, plantains, corn, sugar, and coffee, that are at risk during El Niño events. The region is also home to many subsistence farmers who rely on rainfed agriculture. These climate anomalies can lead to loss of livelihood and regional food insecurity. The team used remote sensing data from Global Precipitation Measurement (GPM) mission Integrated Multi-satellite Retrievals for GPM (IMERG) and Terra Moderate Resolution Imaging Spectroradiometer (MODIS) to generate Normalized Difference Vegetation Index and Standard Precipitation Index data to identify regions that have been negatively impacted by past El Niño events in the dry corridor of Guatemala, Honduras, and Nicaragua. The project identified the historic areas of socioeconomic vulnerability during the onset of El Niño related drought. Team members worked with Universidad del Valle de Guatemala to provide information regarding regional drought to the Nicaragua Ministry of Agriculture and Forestry and the Honduran Ministry of Agriculture and Livestock. The project concluded that there is an increase in the area of severe and extreme drought during El Niño events in the Central American dry corridor, and the team provided time series maps that visually demonstrated the progression of drought in the region.

Advisor: Andrew Shannon (NASA DEVELOP)