Aggressive Flight With Suspended Payloads Using Vision-Based Control

Payload manipulation with aerial robots has been an active research area for many years. Recent approaches have sought to plan, control, and execute maneuvers with large, yet deliberate, load swings for more agile, energy-optimal maneuvering. Unfortunately, the system's nonlinear dynamics make executing such trajectories a significant challenge and experimental demonstrations thus far have relied completely on a motion capture system and non-negligible simplifications like restriction of the system to a two-dimensional workspace or closing of the control loop on the quadrotor, instead of the payload. In this work, we observe the payload using a downward-facing camera and estimate its state relative to the quadrotor using an extended Kalman filter. We demonstrate closed-loop payload control in the full three-dimensional workspace, with the planning, estimation, and control pipeline implemented on an onboard processor. We show control of load swings up to 53
^o
from the vertical axis. To the best of our knowledge, this represents the first realization of closed-loop control of agile slung-load maneuvers and the largest achieved payload angle.