摘要：Waypoint-based trajectory tracking control of a quadrotor UAV with robust performance attributes requires the generation of dynamically feasible and smooth trajectories. The smoothness of these trajectories relies on the adopted interpolation methods along with the constraints being considered (e.g. velocity, acceleration, and input constraints etc.). This paper briefly discusses the optimal trajectory generation and investigates Runge’s phenomenon by presenting various feasible interpolation methods. Runge’s issue may become severe due to sampling between waypoints when corridor constraints are considered with higher ordered polynomial interpolation. Trajectory generation requires the system dynamics to be differentially flat, therefore, a detailed proof is presented as well which shows clearly that quadrotor underactuated dynamics are differentially flat. For corridor constraints, higher order polynomials generate the trajectories near boundary walls, which may result in collision. This problem is solved using a scheme that guides these polynomials from one waypoint to another by incorporating a guiding term. The degree of smoothness and the shape of the desired path can be adjusted through the afore-mentioned guiding term. The proposed trajectory generation scheme can effectively reduce Runge’s phenomenon and still offer finite and smooth curvatures at waypoints. Furthermore, the quadrotor model along with classical linear controller is used to demonstrate the smooth and faithful tracking of the optimal trajectories generated.
关键词：Trajectory generation; Guided polynomials; Quadrotor UAV; Aerial robot; Linear controller
