This thesis presents model-based adaptive motion control algorithms for under-actuated
aerial robotic manipulators combining a conventional multi-rotor Unmanned Aerial Vehicle
(UAV) and a multi-link serial robotic arm. The resulting control problem is quite
challenging due to the complexity of the combined system dynamics, under-actuation, and
possible kinematic redundancy. The under-actuation imposes second-order nonholonomic
constraints on the system motion and prevents independent control of all system degrees
of freedom (DOFs). Desired reference trajectories can only be provided for a selected
group of independent DOFs, whereas the references for the remaining DOFs must be determined such that they are consistent with the motion constraints. This restriction prevents
the application of common model-based control methods to the problem of this thesis. Using
insights from the system under-actuated dynamics, four motion control strategies are
proposed which allow for semi-autonomous and fully-autonomous operation. The control
algorithm is fully developed and presented for two of these strategies; its development for
the other two configurations follows similar steps and hence is omitted from the thesis.
The proposed controllers incorporate the combined dynamics of the UAV base and the serial
arm, and properly account for the two degrees of under-actuation in the plane of the
propellers. The algorithms develop and employ the second-order nonholonomic constraints
to numerically determine motion references for the dependent DOFs which are consistent with the motion constraints. This is a unique feature of the motion control algorithms
in this thesis which sets them apart from all other prior work in the literature of UAVmanipulators.
The control developments follow the so-called method of virtual decomposition,
which by employing a Newtonian formulation of the UAV-Manipulator dynamics,
sidesteps the complexities associated with the derivation and parametrization of a lumped
Lagrangian dynamics model. The algorithms are guaranteed to produce feasible control
commands as the constraints associated with the under-actuation are explicitly considered
in the control calculations. A method is proposed to handle possible kinematic redundancy
in the presence of second-order motion constraints. The control design is also extended to
include the propeller dynamics, for cases that such dynamics may significantly impact the
system response. A Lyapunov analysis demonstrates the stability of the overall system and
the convergence of the motion tracking errors. Experimental results with an octo-copter integrated with a 3 DOF robotic manipulator show the effectiveness of the proposed control strategies. / Thesis / Doctor of Philosophy (PhD)
| Identifer | oai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/22801 |
| Date | January 2018 |
| Creators | Jafarinasab, Mohammad |
| Contributors | Sirouspour, Shahin, Electrical and Computer Engineering |
| Source Sets | McMaster University |
| Language | English |
| Detected Language | English |
| Type | Thesis |
Page generated in 0.0019 seconds