Jake Welde

Jake Welde

PhD candidate in robotics working at the intersection of geometry, dynamics, and design.

Task Space Planning for Underactuated Aerial Manipulators

2021

with James Paulos and Vijay Kumar

A simulated aerial manipulator, consisting of a quadrotor equipped with a 2-DoF manipulator arm, performs a 6-DoF manipulation task precisely, by jointly leveraging the actuators in the arm and in the vehicle while respecting the underactuation constraints of the combined system.

In this work, we address the problem of planning dynamically feasible trajectories for underactuated aerial manipulators to achieve a desired trajectory for the end effector. We consider a quadrotor equipped with an arbitrary n-joint articulated arm. We show that the combined underactuated system is differentially flat, however the flat outputs do not correspond directly to the motion of the end effector. We therefore develop a method which determines the family of flat output trajectories which will exactly produce any desired task trajectory, even in the case of dynamic maneuvers. We also give criteria on the manipulator geometry which will ensure certain important stability properties, informing hardware design. The entire approach is demonstrated in simulation for systems of varying geometry and number of joints. The simultaneous resolution of the kinematic and dynamic constraints allows these tasks to be performed dynamically without sacrificing accuracy.


Relevant Publications

  1. "Coordinate-Free Dynamics and Differential Flatness of a Class of 6DOF Aerial Manipulators",
    Jake Welde and Vijay Kumar.
    IEEE International Conference on Robotics and Automation (ICRA), 2020.
  2. "Dynamically Feasible Task Space Planning for Underactuated Aerial Manipulators",
    Jake Welde, James Paulos, and Vijay Kumar.
    IEEE Robotics and Automation Letters (RA-L), 2021.
    Finalist for "Best Paper in Unmanned Aerial Vehicles" at ICRA 2021.