Exploiting Image Moments for Aerial Manipulation Control

Abstract

This paper proposes a new visual servo control scheme that endows flying manipulators with the capability of positioning with respect to visual targets. A camera attached to the UAV provides real-time images of the scene. We consider the approaching part of an aerial assembling task, where the manipulator carries a structure to be plugged into the visual target. In order to augment the system capabilities regarding the 3D interaction with the target, we propose to use image moments. The developed controller generates desired velocities to both the UAV and the manipulator, simultaneously. While taking into account the under-actuation specific to rotary-wing vehicles, it makes use of the system redundancy to realize potential sub-tasks. The joints limits avoidance is also guaranteed. The presented developments are validated by means of computer simulations.

Passivity-based control of VToL UAVs with a momentum-based estimator of external wrench and unmodeled dynamics

“…, 3. Notice that with small modifications to (12), it is possible to reach a transfer function in (10) of the desired order.…”

“…Under the given assumptions, considering the dynamic model of a generic VToL UAV for the angular part (4b), the designed control law (15) and the compensation of the estimated moments in (12), the state error x 2 , whose dynamics is given by the closed loop system (22b), is ultimately bounded as…”

“…with  f u given by (12), while the desired yaw ψ d is given by the planner. A secondorder low-pass digital filter should be employed to reduce noise and compute both first and second derivatives of η d [17], so as to getη d andη d , and hence compute in turn the attitude tracking errors e η andė η .…”

“…A secondorder low-pass digital filter should be employed to reduce noise and compute both first and second derivatives of η d [17], so as to getη d andη d , and hence compute in turn the attitude tracking errors e η andė η . The control input vector τ b b is computed as in (15), in which  τ u is given by (12). Having both the thrust u and the actuation torques τ b b , the squared propellers speeds w 2 i of the VToL UAV, with i = 1, .…”

“…From passive tasks like inspection [1,2], surveillance and monitoring [3], remote sensing and so on, such aerial vehicles are now beginning to be employed in active tasks like grasping [4] and manipulation [5][6][7][8][9][10][11][12][13][14][15][16]. This change of perspective requires the UAV to operate in changing and unstructured scenarios.…”