Decision Making on AUVs for adaptive minehunting surveys

Minehunting operations can benefit from the use of Autonomous Underwater Vehicles (AUVs)
because these vehicles are able to acquire high-resolution images of the seabed leaving the man out of the
minefield. However, because there are severe communication constraints between the AUV and the
control platform, the AUV needs to be capable of changing its behaviour when situations are encountered
that differ from those assumed in the mission planning. Therefore, the AUV should have a capability to
autonomously adapt to changing circumstances. This capability is important at different time scales
during a mission, ranging from quick reactive responses to avoid objects to deliberative responses relating
to the mission. In this paper, the focus is on the deliberative response at mission level. The specific case
evaluates whether sufficient information is collected in the search-classify-map (SCM) phase to fulfil the
mission requirements on effectiveness and efficiency. Based on this evaluation, decisions are made to
extend the SCM-phase, to start the reacquisition and identification, or to abort the mission when the area is
considered to be too complex for autonomous minehunting. The challenge is to enable the system to
reason about measures of effectiveness such as the achieved detection performance in the mission. To
enable this, the system has to be able to compute the value of the collected information using in-situ
observations. Subsequently, it needs to choose the best future course of action taking into account the
available resources. The evaluation of measures of performance and effectiveness in-mission and the
subsequent decision-making has been demonstrated on the MUSCLE AUV during the MANEX’13 trial
organised by the Centre of Maritime Research and Experimentation (CMRE).