Underwater Robots - Gianluca Antonelli.pdf

226 9. Coordinated Control of Platoons of AUVs
is not achieved by resorting to deterministic, pre-programmed, paths but is
made adaptive in order to extract the maximum possible information during
one mission. The formation control is obtained using virtual bodies and
artificial potential techniques [187, 218]; the Authors separate the problem
into small area coverage (5 km) and synoptic area coverage for larger scale
(5–100 km). Forthe former, experimental result performed with 3gliders in
2003 are reported; the experiments were successful thus demonstrating the
reliability ofthe approach inareal environment under the effect, e.g, of the
ocean current.
In [54] acooperative mission, whose experimental validation is planned
for summer 2006, isdescribed. The mission objective istoperform aside-scan
survey searching for mines with mixed initiative interactions; the presence of
ahuman operator inthe loop, thus, is considered. In detail, several target
mines will be deployed inabay and the vehicles have to find and classify the
targets inminimum-time.
In [275, 276], an effective decentralized control technique for platoons is
proposed and simulated for underwater vehicles. Remarkably, the approach
requires alimited amount ofinter-vehicle communication that isindependent
from the platoon dimension; moreover, control ofthe platoon formation
is achieved through definition of asuitable (global) task function without
requiring the assignment of desired motion trajectories to the single vehicles.
In Figure 1.7, one ofthe AUVs developed at the Virginia Tech is shown,
these vehicles are very small in size and cheap with most of the components
custom-engineered.
Reference [179] presents abehavior-based intelligent control architecture
that computes discrete control actions. The control architecture separates the
sensing aspect, called perceptor in the paper, from the control action, called
response controller and itfocuses onthe latter implemented by means of a
set of discrete event models. The design of asampling mission for AUVs is
also discussed in the paper.
In [158], the problem of formation control for AUVs is solved by using a
leader-follower approach. Foreach vehicle arelative position with respect to
the leader is given as areference motion; simulation for two 3-dimensional
nonholonomic vehicles are reported.
Reference [55] presents an algorithm to perform asearch using multiple
AUVs. Simulation results for the search ofthe minimum in ascalar fields are
shown.
In [247] the communication issue for the underwater environment istaken
into account. The vehicles operate in decentralized manner and communicate
to achieve common control objectives. The network, thus, is time-varying,
and the vehicles communicate with different subset of other vehicles along
the mission. In particular, the case of disconnected network is addressed and
possibly solved by resorting to a fast enough periodic fast switching. The