Pilot Case 5
Robotic harvesting in orchards in Belgium
Location: Belgian Orchard farmers and cooperatives
Overview
Pilot Case 5 develops an autonomous orchard robot equipped with a manipulator and soft vacuum gripper for selective fruit harvesting. The robot is built on a mobile base (as shown in the render) instrumented with LiDAR, 3D cameras, IMUs, and GNSS for autonomous navigation. This base moves from tree to tree, carrying a 2D gantry that shifts vertically and sideways. Mounted on the gantry is a 6-degree-of-freedom industrial arm that can reach branches and fruit within a semi-spherical workspace.
At the end of the arm sits the vacuum gripper and a tool-mounted 3D camera, allowing the system to identify and gently harvest fruit with minimal damage. The combination of GNSS and vision-based navigation ensures accurate positioning while also computing real-time yield statistics during harvest operations.
AgRibot orchard robot platform for Pilot Case 5 — autonomous navigation with LiDAR, cameras, and GNSS, equipped with a 6-DoF arm and soft vacuum gripper for selective fruit harvesting.
Timeline
An iterative pilot, starting with multi-season data collection and fruit detection model training. Field trials will validate navigation and picking performance, with robot components refined between testing rounds based on user feedback.
Validation
The pilot evaluates: (i) navigation accuracy, (ii) fruit detection precision, (iii) picking efficiency and fruit integrity, and (iv) overall harvest speed and productivity. Integrated performance will be tested in real orchard conditions.
AR Integration
AR highlights ripe fruits for robotic harvesting, while XR simulations train new users in orchard layouts, robotic operation, and troubleshooting—bridging the gap between traditional and tech-augmented farming.