Affiliations: [a] Computational Intelligence Group of the University of the Basque Country, UPV/EHU, Bilbao, Spain | [b] ENGINE Centre, Wrocław University of Technology, Poland | [c] ACPySS, San Sebastian, Spain
Correspondence:
[*]
Corresponding author: Julian Estevez, Computational Intelligence Group of the University of the Basque Country, UPV/EHU, Bilbao, Spain. E-mail:[email protected]
Abstract: The aim of this work is to design robust control algorithms
of aerial robots, i.e. quadrotors, for team transportation of a deformable
linear object (DLO). The DLO-robot attachment makes
the whole system physically coupled by an elastic
element, which introduces strong non-linearities
in the system dynamics. Sections in quasi-stationary state of the
DLO hanging freely from two extreme points can be modeled by catenary
curves, so we are able to build a detailed and accurate simulation
of the system as the workbench for the evaluation
of alternative quadrotor control approaches. DLO
transportation is an instance of a follow the leader team strategy,
in which the local quadrotor control must cope with the dynamic perturbations
due to the DLO linking the quadrotors. The quadrotor team control
has two phases, one achieving a spatial configuration with equal energy
consumption, the other is to manage the horizontal motion which is
the transportation process per se.
In this paper we deal with the second control problem, assuming the
first solved. We use proportional derivative (PD) controllers for
both quadrotor attitude and trajectory control. Offline
PD parameter setting carried by metaheuristic optimization can not
cope with the perturbations induced by the DLO and environmental conditions,
i.e. wind shear. Therefore we propose an adaptive controller for
the quadrotors carrying out the DLO transportation task which uses
fuzzy modeling of the error in order to modulate the activation of
the PD parameters adaptation rules, which are error
gradient descent rules. Computational experiments on our system simulation
workbench show that our adaptive approach scales well when increasing
the number of quadrotors, providing smooth follow-the-leader team strategy navigation
behaviors even under strong wind perturbation conditions. We
compare our approach with other state-of-the-art online and offline
approaches.
