A gentle and versatile robotic gripper for efficient crop harvesting

Robotic grippers have become essential across many industries, including manufacturing, packaging, and logistics, mainly for pick-and-place tasks. Recently, the demand for robotic grippers has also expanded into agriculture, where they are used for harvesting and packaging tasks.

However, conventional robotic grippers struggle with the unique shapes, properties, and delicate nature of different crops. Consequently, there has been an increasing demand for more versatile robots that can adapt to objects with various shapes, sizes, and textures.

Robotic grippers that are made of soft materials have emerged as a potential solution to the above problem. However, current methods for adapting these grippers to complex geometries rely on complex control and planning generated by data-based models.

These models require a large amount of data, limiting their general applicability. Additionally, integrating a sensory system into their soft body requires complex designs and sophisticated fabrication methods.

To this end, a team of researchers from the Japan Advanced Institute of Science and Technology (JAIST), led by Associate Professor Van Anh Ho, along with Assistant Professor Nguyen Huu Nhan and doctoral course student Nguyen Thanh Khoi, developed an innovative soft robotic gripper named ROtation-based Squeezing grippEr or ROSE.

“ROSE takes inspiration from the blooming states of a rose to generate grasping action. It offers a simpler approach to real-farm harvesting by gently grasping objects using a unique ‘wrinkling’ phenomenon.

“Unlike conventional grippers, ROSE doesn’t require complex control and planning strategies to adapt to various agricultural products with diverse shapes, sizes, and textures,” explains Prof. Ho. The team also employed a simulation model to fully understand and optimize the grasping mechanism of ROSE.

The study is published in the International Journal of Robotics Research.

ROSE consists of an isolated cup-shaped chamber formed by two thin, soft elastomer layers, with a separation between the interior and outer layers. Rotating only the inner layer using an external motor produces a deformation in the layers.

Specifically, this twisting motion of the inner layer results in a strain mismatch between the outer and inner layers, resulting in the formation of a series of wrinkle-like inward folds, a process termed “wrinkling.” This unique mechanism shrinks the central space in ROSE, which allows it to gently grasp any object present within this central space.

To refine this mechanism, the researchers studied the wrinkling process through a finite element method-based simulation model. The simulations revealed a correlation between different geometrical features, including thickness, diameter, and height. Notably, they found that an appropriate distribution of ROSE’s skin thickness, that is, the separation between the layers, has a significant influence on its grasping performance.

As a result, the researchers tested two different thickness distribution strategies, namely linear and non-linear distribution, which significantly improved ROSE’s grasping performance compared to a constant thickness. Moreover, the simulations also highlighted the importance of the ratio between the gripper‘s diameter and height.

The simulation results were validated through various experiments, verifying ROSE’s capability to carry out tasks that are difficult for traditional grippers.

The researchers demonstrated the practical applications of ROSE in agriculture by using it to harvest strawberries and mushrooms. ROSE achieved high success rates in picking up these crops in multiple trials, regardless of whether they were stiff or soft. It also succeeded in picking up a clump of mushrooms without breaking any pieces, provided that the clump size fit within the grasping space.

“ROSE is one of the first grippers to utilize buckling deformation as a gripping method, challenging the conventional mindset that the buckling phenomenon is an undesired feature. Moreover, the practical application of ROSE in agricultural settings is a game-changer for real-farm harvesting.

“ROSE’s ability to adapt to varying textures and shapes makes it highly effective in these tasks. This not only improves efficiency but can also address the growing labor shortages in agriculture, particularly in regions with aging populations,” states Prof. Ho.

This study marks a significant advancement in robotic grippers. With gentle yet adaptable grasp positions, ROSE might usher in a new era of automated farming.