research 6-DOF eVTOL Simulation & Adaptive Control Nonlinear simulation with MRAC, L1, LPV, Geometric, LQR, and H-infinity controllers validated through Lyapunov analysis and Monte Carlo studies Numerical Optimal Control of a UAM Vehicle Hermite-Simpson direct collocation with projected gradient descent for trajectory optimization of a full 12-state nonlinear quadrotor model — AOE 5404, Spring 2025 DRAGON eVTOL — Fixed-Wing to Multirotor Transition PX4 firmware modification, gain-scheduled LQR/PID transition control laws, and first demonstration of stable fixed-wing-to-multirotor flight Gradient Descent eVTOL Design Optimization Custom gradient descent with penalty method and BEMT-based propulsion model to minimize takeoff weight over wing area, rotor count, diameter, pitch, and cruise velocity ROS 2 GNC Middleware with EKF Sensor Fusion Low-latency C++/Python ROS 2 middleware for time-synchronized multi-sensor acquisition with Extended Kalman Filter state estimation L1 Adaptive Control — Race Drone Testbed Custom race drone built as a high-bandwidth hardware testbed for validating L1 adaptive augmentation on a real agile platform Autonomous Path Planning — Scratch-built Autopilot End-to-end autonomous path planning implementation on a custom-built autopilot stack industry 500 lb Heavy-lift Multirotor UAV End-to-end technical ownership from motor and frame sizing through avionics integration, GNC commissioning, and first autonomous flight 1 kHz Multi-channel Propulsion Test Stand NI DAQ-based dynamometer capturing thrust, torque, RPM, phase current, and winding temperature at fidelity exceeding commercial dynamometers at the required scale Physics-based eVTOL Propulsion & Aerodynamic Models MATLAB state-space propulsion models, motor torque-speed curves, battery discharge models, and coupled airframe aerodynamics for eVTOL control law development other Rolls-Royce Viper Thermal Failure Investigation Aero-thermal failure analysis of the central bearing, identifying heat transfer breakdown at the bearing race as the primary failure driver
