Subscribe to our newsletter to receive the latest news and events from TWI:

Subscribe >
Skip to content

SoftGrip Robotic Grippers Imitate Human Harvesting Skills

Wed, 29 September, 2021

In early 2021, TWI Hellas joined a prestigious European consortium – comprising SSSA (Italy), ICCS-NTUA (Greece), UESS (UK), TEAGASC (Ireland) and MITSUI (Germany) – to jointly work on a project to create autonomous robotic grippers that mimic human harvesting skills.

The robotic grippers being developed will learn, test, and polish their skills on the white button mushroom, which is the most prevalent fungus variety globally. Picking mushrooms is challenging because they can easily be blemished or damaged by excessive pressure and unsuitable rotating movements. This is also the reason that limited robotic systems have been built and implemented in the mushroom farming industry thus far.

Project SoftGrip embraces this challenge and derives its strength from the multifaceted team of experts in functionalised structure fabrication, soft robotics design and control, artificial intelligence, mushroom cultivation and provision of prototypes. TWI Hellas, as the project partner with expertise in artificial intelligence (AI), robotics, machine learning and automation, is focusing on control system integration and validation, including the grasping control of the grippers, contact-rich simulation and finite element analysis (FEA).

According to a recent report, the global robot gripper market is estimated to reach $1.46 billion by 2027 (accessed 27 September 2021). These devices will be of utmost value to several industries, in particular to soft robotics – the subset of robotics that seeks to resemble the physical capabilities of a living organism as closely as possible, which will witness a significant rise in agriculture – as well as materials handling, pharmaceuticals, food and beverage, industrial machinery, logistics and others.

SoftGrip, as its name suggests, intends to perfect the ‘soft touch’ when handling delicate food like mushrooms, kiwis and berries. The project employs a learning-by-demonstration approach in order to build up the knowledge base of the robotic grippers, and accelerate skills transfer by implementing machine learning algorithms and methodologies. To achieve this, the multidisciplinary team behind the project is concentrating on developing intelligent, soft gripping devices that have the inherent ability to perform in a distributed and continuous manner. Behind this, lies a meticulous study into the force and torque movements that occur during the picking process, which is enabled by harvesters wearing smart gloves in order to register their every movement.

The consortium members will consider different aspects of the end-to-end handling process, alongside the effective simulation of gripping movements, to ensure that the soft grippers are food-safe and able to self-heal, enabling them to remain effective after thousands of operating cycles, as well as able to overcome microscopic and macroscopic damage while not affecting the food being picked in any way. In addition, they are intended to be recyclable when they reach end-of-life.

Longer term, the SoftGrip, European initiative looks to the future in its aims, with an overarching goal to implement the technology in various industries, including the soft fruit market but not limited to agriculture, in order to bring about technological change that will revolutionise present-day, complex harvesting processes.

To find out more about the Horizon2020 project, visit the SoftGrip website.

Athanasios Mastrogeorgiou, a Robotics Software Engineer at TWI Hellas, discussed the company’s contribution to the project saying “We will employ our extensive, practical experience in robotic systems to bring to life robotic grippers leveraging state of the art 3D object segmentation algorithms and imitation learning techniques, for the effective harvesting of mushrooms and soft fruit.”

For more information please email: