In Underactuated Robotics, I and a fellow student developed a project to plan safe kinematic paths through cluttered space with an optimal Rapidly-exploring Random Tree algorithm. The paths were then fed to a trajectory optimization on the quadrotor to account for dynamics, further obstacle avoidance constraints, and to find optimal input and state trajectories for the quadrotor. These trajectories were then simulated within Drake and stabilized using finite-horizon LQR. Take a look at our video presentation and final report below!