Towards Autonomous Inland Shipping.

Road transport dominates the European and Belgian hinterland freight transport
sector. Over the last decades, approximately three quarters of the cargo
streams passed by road, whilst inland waterway transport seems to offer a more
sustainable alternative. Moreover, the larger inland vessels suffer from an excess
of supply over demand, and the amount of smaller inland vessels is diminishing.
This negative evolution in the inland waterway transport sector does not align
well with the European and Belgian governmental ambitions to transfer more
freight transport to their waterways. This thesis aims to investigate a solution for
this emerging tension field by studying the technological feasibility of unmanned
inland cargo vessels. More precisely, three research questions are answered.

The first question explored: “How to design and construct an industrially
relevant research vessel for unmanned inland cargo shipping?”. To investigate
the industrial relevance, the present inland waterway transport sector was
examined. Three developments stood out. First, the European Watertruck+
project introduced a novel fleet of modular push boats and barges. Hence, the
watertrucks can separate their navigation time from their cargo handling time.
Second, Blue Line Logistics built new flat deck vessels with an onboard crane,
which enables them to handle their cargo independently from the shoreside
infrastructure. These vessels focus on transporting palletized cargo. Finally,
smaller urban freight vessels have already successfully transported cargo within
several European cities. Next to these industrial developments, the recent
research evolution regarding unmanned shipping in general and the specific
challenges for inland waterway transport motivated the build of two unmanned
research vessels: a scale model self-propelled watertruck barge and a functional
scale model of a flat deck barge which focusses on palletized cargo. In addition,
both research vessels have a length that facilitates intracity freight transport
research.

The second question investigated: “How to model and identify the hydrodynamic
motion models of an inland cargo vessel?”. The modelling part focussed mainly
on the decoupled equations of motion in the water plane, i.e., the surge, sway,
and yaw degrees of freedom. To identify this decoupled model, experimental data were fetched with the research vessel in its real outdoor environment. Two
identification procedures were compared. The first one used the instantaneous
force balance, and the second one integrated the differential equations of the
decoupled motions. Furthermore, two independent data sources were used
to validate the identified models: bollard pull test data, measured inside a
towing tank, and longitudinal damping data, calculated via computational fluid
dynamics.

The third question studied: “How to provide an unmanned inland cargo vessel
with perception and motion control?”. For this purpose, four navigational
environments were differentiated, based on the presence of known or unknown
and static or dynamic objects. These environments influence the requirements
for the perception and motion control systems of the vessels: exteroceptive
sensors are needed to detect unknown objects, and traffic rules need to be
implemented in order to avoid dynamic objects. This thesis demonstrates the
first successful missions of an unmanned and autonomous vessel navigating on
a river with known static obstacles.

Furthermore, this work provides an alternative answer for the last two research
questions, by the construction of an inland shore control centre to remotely
monitor or control vessels. Accordingly, the operator performs the perception
and motion control tasks for the vessel, and implicitly models and identifies
the behaviour of the ship. A shore control centre, however, raises new research
questions: can this centre help the operator to construct a feeling of ship sense,
and can the operator keep the ship in harmony with the environment from
a remote location? The initial experiments, with an operator in this control
centre remotely controlling an unmanned vessel, delivered a first answer for
these novel questions. In addition, this thesis includes some preliminary results
with an augmented remote control system in the control centre. This augmented
system offers the operator extra visualisations and measurements of the vessel
on its navigational chart.

Evidently, the technological feasibility of the abovementioned research questions
alone cannot judge the socio-economic feasibility of unmanned inland shipping
in general. Consequently, this work aims to gain insights in order to enable
higher resolution socio-economic feasibility studies, with the ambition to guide
the course of future investments streams.