An ellipsoidal-polytopic based approach for aggressive navigation using nonlinear model predictive control

This work addresses the development of a nonlinear model predictive controller (NMPC) for controlling autonomous motion systems navigating within cluttered environments with unknown obstacles. More specifically, this work concerns the aggressive maneuverability of motion systems, allowing them to navigate at high speed while reacting dynamically to obstacles. Additionally, these applications require the ability to generate and control large linear and angular accelerations. In this regard, a new ellipsoidal-polytopic approach for obstacle avoidance is incorporated by the NMPC. As a result, the new control method considers the orientation of the motion system in the collision avoidance, allowing aggressive maneuvers to improve the motion system performance for the obstacle deviation. The new controller is corroborated with a quadrotor UAV, which accomplishes a point-to-point motion avoiding obstacles positioned between the starting point and the destination point.