High-Fidelity Robot Simulation
Experimental Validation of Simulation
Navigation, flight, and landing dynamics were all built from initial conception, tested in a custom Gazebo simulation, and validated with physical hardware in experimental tests.
System Identification and Modeling
System inertia was estimated through bifilar pendulum method. And custom modeling of motor behavior was performed through use of thrust stand testing.
Simulation Architecture
Simulation was built using modern C++ and Python system design principles, combining multiple concurrent processes and ROS nodes. These processes would coordinate and manage communication between ML training, data logging, robotic control, and system modeling.
