Kripper is a kirigami shell gripper for food production lines. One cut-and-folded sheet closes into an envelope around a tomato, a croissant or a pepper, senses how well it is holding on, and adjusts before anything slips or bruises.
Robot arms, vision and controls are now off-the-shelf. The tool at the end of the arm is what stops delicate produce from being automated.
of total costs in Dutch greenhouses goes to labour, about €300,000 per company per year. [1]
seasonal workers a year keep EU horticulture running, often in physically demanding conditions. [2]
EU tomato production value in 2022, 19.1% of all fresh vegetable production. [3]
of packhouse throughput is lost to handling damage: bruising, downgrade and rejection.
Suction cups lose their seal on wet, curved or floured surfaces. Soft pneumatic fingers run open-loop at a fixed pressure. Neither knows whether it is actually holding the product.
| Suction / vacuum | Soft pneumatic fingers | Kripper | |
|---|---|---|---|
| Needs | Smooth, dry surfaces to hold a seal | Product inside the tuned size and stiffness range | An object inside the shell's envelope. Size, shape and stiffness can vary. |
| Fails on | Moist, curved, porous or floured surfaces | Fruit that varies outside the envelope it was tuned for | Designed for the cases on the left |
| Infrastructure | Vacuum generation, energy-hungry and noisy | Compressors, air lines, valves and tubing | Electric single-axis actuator, no air |
| Sensing | Vacuum level, a proxy for grip | Little to no contact sensing | Pressure matrix, temperature and moisture at the contact surface |
| Hygiene | Cup crevices are hard to clean | Membranes and folds trap residue | One monolithic sheet, no hinges or cavities |
Kirigami is the Japanese craft of cutting and folding paper. In engineering, a pattern of cuts in a thin shell turns one simple input into a coordinated 3D movement, with no hinges, joints or linkages. The same principle already folds space structures and medical devices.
The shell closes into an envelope around the object, so it holds without a pinching force. It can also grip with its edges where an envelope does not fit.
The shell deforms when it closes. It secures the grip by giving way, not by pressing harder.
One shell covers a range of sizes, shapes, stiffness and presentation. Seasonal products do not need a tooling change.
A single part with no moving joints. Cleanable, light and inexpensive to replace.
Vision finds the object and sets the approach and the shell's pre-shape before contact.
Closing force is adjusted in real time from the tactile skin, within hard force limits.
When grip quality drops, a learned policy adjusts contact force to recover before the product slips.
Each attempt is logged with object, forces and result, and the grasp policy improves from it.
Retrain, don't retool. A new crop or product family is added by training on new data. The gripper stays on the arm.
A thin flat shell makes an ideal base for sensors. The sensor stack bonds straight onto the surface, with no housings or cavities.
Picking, re-orienting and placing individual fruit into trays. On moist, curved skin, suction seals fail and rigid jaws bruise. Validated first on tomatoes, then extended to peppers and cucumbers by retraining.
Placing rolls, buns, pastries and loaves into trays and cartons. Floured, sugared crusts will not hold a vacuum, and any clamping force marks the crumb. PPWR is moving the sector to cardboard formats that need gentle placement.
Lab validation runs blocks of more than 1,000 grasp attempts against suction and soft pneumatic baselines on the same objects. The field pilots then benchmark Kripper against the incumbent tool on a real line.
We are looking for packhouses, bakeries and integrators who want to benchmark Kripper against their current tooling. Tell us what you handle and where it goes wrong.
Ambionic B.V. Bakkerstraat 39 A 3082 HB Rotterdam The Netherlands