Research
Research
Master's Thesis
Vision Based Hazard Detection & Real-Time Trajectory Optimization via Successive Convexification for Lunar Landing
With the renewed focus on lunar exploration, ensuring safe and autonomous lunar descent is crucial for mission success. This research presents a framework for hazard detection and real-time trajectory optimization during the final stages of lunar landing. Using computer vision, hazards like craters and mountains are identified from descent imagery and mapped with fused altimetry data. The lander then optimizes its trajectory by solving a receding-horizon control problem that incorporates chance constraints to account for uncertainty in hazard mapping and dynamics. Successive convexification enables efficient computation of these safety-constrained, minimum-time trajectories, ensuring feasibility in real time under fuel and maneuverability limitations. Simulations in a 3D lunar environment validate this adaptive framework for future missions.
Proposal Announcement: LinkThesis Defense: April 1st (Tentative)Source: LinkFaculty Advisor: Dr. Kyriakos Vamvoudakis
Proposal Announcement: LinkThesis Defense: April 1st (Tentative)Source: LinkFaculty Advisor: Dr. Kyriakos Vamvoudakis
Undergraduate
Urban Air Mobility: Motion Planning
Fall 2023
Faculty Advisor: Dr. Kyriakos VamvoudakisSkills: Python, JuliaSource: link
- Developed Actor-Critic Reinforcement Learning for roll, pitch, yaw and thrust control for COTS drones.
- The drones were used for Decentralized motion planning in urban areas.
Faculty Advisor: Dr. Kyriakos VamvoudakisSkills: Python, JuliaSource: link
Icesat-2 Data Analysis
Fall 2021 - Fall 2022
Faculty Advisor: Dr. Brian GunterSkills: Python, Linux Bash Source: link
- Performed Data Reduction of Icesat-2 LIDAR Land-Ice levels using weighted Averaging on PACE Cluster.
- Performed PCA and and determined EOF for spatial and temporal land-ice analysis.
- Analysis on cross-covariance of Land-Ice and GRACE gravimetry data.
Faculty Advisor: Dr. Brian GunterSkills: Python, Linux Bash Source: link
Mars Aerobraking (Virginia Tech)
- Conducted Aerobraking Background research and simulated 14 aerobraking drag passes in MATLAB.
- Analyzed heat flux, drag trends and reported results and further developments.
Faculty Advisor: Dr. Kevin Schroeder Skills: MATLAB