Light Curve Classification with Machine Learning Nat Sutton, Daniel Warshofsky, Michael Coughlin, Jan van Roestel Project Mentor(s): Darci Snowden, PhD; Cassandra Fallscheer, PhD
Stars are classified by their light curves, a representation of their luminosity over time. These classifications are used in the study of stellar objects to characterize their behavior. This project used machine learning techniques to identify eclipsing white dwarf binaries in a large data set collected by the Zwicky Transient Facility. This is important because classifying this data by hand is time consuming and error prone, so this project aimed to reduce the categorization needed down to a manageable level. From a 10 thousand item data set, a decision tree was used to classify white dwarf binaries, with a final accuracy of over 90%. The next steps are to modify the code to run over a database instead of a select set of data. Presentation Type: Poster Presentation (May 21, 9:30am–3:00pm) Keywords: Variable stars, Astronomy SOURCE Form ID: 127 Project Orchid Mantis Algirdas Veitas, Erik Hardiman, Dylan McGovern, Jace Leensvaart, Carter Shafer, Brandon Greenleaf, Armondo Nungaray, Matthew Thien, Jairus Phillips, Colin Thomas, Desmond Hill † , Jose Vasquez- Sanchez Project Mentor(s): Darci Snowden, PhD; Tammy King We focuse on design, fabrication, integration, and flight testing of a high power rocket developed for the International Rocket Engineering Competition. The rocket is being built as an instrumented platform capable of carrying multiple payload systems to an altitude of 10,000 feet while collecting useful flight data. The project combines structural design, propulsion, avionics, recovery, payload integration, and control systems into one complete vehicle. The rocket uses an aluminum fin-can and reinforced internal structure to withstand launch, aerodynamic, and recovery loads while maintaining structural integrity. Propulsion is provided via commercial solid rocket motor selected to safely launch the vehicle and provide enough total impulse for the mission. Redundant flight computers, independent power systems, and GPS tracking are included to improve reliability and simplify recovery. A dual deployment parachute system controls descent. The vehicle carries sensors that record altitude, acceleration, vibration, position, and onboard video throughout flight. Additional elements include CubeSat landing device, ejection mechanism for payload deployment, and a vibration dampening system to protect sensitive electronics and improve survivability. Another component of the project is an active airbrake system designed to control apogee and reduce altitude variation caused by changing conditions. Flight data will be used to compare predicted and measured performance, evaluate control effectiveness, and guide future improvements.
Presentation Type: Poster Presentation (May 21, 9:30am–3:00pm) Keywords: Aerospace, Engineering, Electronics, Mechanical, Physics SOURCE Form ID: 223
143
Made with FlippingBook interactive PDF creator