Planning Optimal Base Positioning for Autonomous Excavators in Olive Harvesting Tasks

Italian olive-growing tradition is going through a difficult period which can cause serious damages to the agricultural sector. Main causes are related to economics and social factors such as high management costs and lack of manpower, due to high risk of injuries and absence of generational change, since younger generations are not interested in this field. The Tuscany region-funded project Robolio aims to tackle these issues by developing a fully autonomous robotic system for olive harvesting. This solution is based on a tracked excavator equipped with an olive harvester with elongated rubber-coated tines as end-effector. The main task of the robotic system is to perform the harvesting of olives in an efficient way, paying attention to the health of the tree. In this context, this thesis, realized in cooperation with Yanmar R&D Europe S.R.L., focuses on the optimization of the excavator’s mobile base placement for olive harvesting, The method is based on the Reachability Map Inversion approach. Specifically, as the first step, a Direct Reachability Map (DRM) is created through an exhaustive Forward Kinematic exploration to get an approximation of the reachable workspace. This spatial map is enriched by adding information about the manipulability of the excavator. Secondly, the DRM is inverted to compute an Inverse Reachability Map (IRM). Finally, this is used to generate an Oriented Reachability Map (ORM), which is a distribution in SE(2) of potential 2D poses of the mobile base that allow reaching a given olive-harvester pose. This leads to finding an optimal (in terms of manipulability) base pose for executing olive-harvesting trajectories. The approach has been successfully evaluated in simulation and experiments using the Yanmar excavator Vio33.