With the growing population and climate change, agricultural productivity growth is unlikely to meet the increased demand for food. Besides the increasing pressure to produce more, there is an overall need for higher quality and sustainable cultivation. Precision agriculture combined with intelligent robotic technologies can push to that direction. The incorporation of such technologies into agricultural production not only benefits productivity but also improves the working conditions of farmers and labourers. Intelligent systems are becoming the go-to solution to push towards precision agriculture, while a large number of farmer operations are already transitioning to full autonomy. Smart, automated and selective harvesting, in particular, can provide considerable improvement in production leaving the unripe product in the field to mature. However, in order to achieve such automation significant progress is required regarding the cognitive and mechatronic capabilities of the robotic agents replacing the human workers in these tasks, especially in cases where human-like actions are required by the robots. BACCHUS intelligent mobile robotic system promises to reproduce hand harvesting operations, while at the same time take the manual legwork out by autonomously operating in four different levels: i) performing robot navigation with quality performance guarantee in order to inspect the crops and collect data from the agricultural area through embedded sensorial system; ii) performing bi-manual harvesting operations with the needed finesse using a modular robotic platform, iii) employing additive manufacturing for adjusting the robot gripper to the geometry of the different crops, iii) presenting advanced cognitive capabilities and decision making skills. The envisioned system will be demonstrated and evaluated in the challenging vineyard environment inspecting different types of vines and harvesting bunches of grapes of different varieties in a human-like manner.