This thesis work aims to develop a Motion Planning Framework that takes into account uncertainty in the robot's pose and configuration for an agricultural mobile manipulator focused on automatic table grape harvesting in Vineyard Environments. The proposed framework consists of two main components: environment reconstruction and motion planning.
Firstly, the environment in the grape's surroundings is reconstructed using Next Best View, an algorithm that scans the environment iteratively and selects the most relevant view. The amount of information provided by a view is estimated using a novel definition of Information Gain, which considers the distance between the target and the camera view and the potential unknown volume that the camera can discover.
The second component of the proposed framework is motion planning. It uses a modified version of ABIT* that tracks the expected uncertainty of the manipulator's pose throughout the path. The uncertainty is projected in task space as a hyper-ellipsoid defined from a χ² distribution that contains the possible poses of the considered point with a given probability. The hyper-ellipsoid is then used as a safety margin in collision checking performed by the planner.
This work extends the idea of uncertainty-aware motion planning for manipulators to realistic work scenarios, enabling the planner to deal with complex environments thanks to the innovative definition of the pose uncertainty in task space. These two components, in combination, allow the planner to adapt easily to a variety of situations, empowering the system with a high level of autonomy.
In conclusion, the developed framework is tested in a simulation environment that resembles a real vineyard, demonstrating its effectiveness and potential for implementation in an autonomous table grape harvesting system in Vineyard Environments.

Keywords: Motion Planning Framework, Uncertainty-aware Planning, Agricultural Robotics, Vineyard Environments, Table Grape Harvesting, Next Best View

The aim of the analysis was to assess the safety of an hydroelectric powerplant.
The analysis made use of multiple techniques: Functional Analysis and operating conditions, PHA (Preliminary Hazard Analysis), FMEA (Failure Mode and Effect Analysis), FTA (Fault Tree Analysis), ET (Event tree).