Project description
Innovative soft gripper for delicate harvesting
Labour costs often represent up to 50 % of total production costs in the fresh food industry. Although the sector demands solutions to reduce them, the shift towards robotic automation is obstructed by the complex and contact-rich interactions needed. The EU-funded SoftGrip project will introduce a self-actuating soft gripper for the autonomous picking of delicate white button mushrooms. The project aims for low-cost, intelligent soft robotic grippers with embedded actuation, tactile sensing, recyclable materials and advanced fabrication techniques. It will develop a set of fast-computed modelling algorithms to enhance real-time model-based control schemes and advanced learning capabilities. SoftGrip will develop a learning-by-demonstration framework that will allow the robot to capture human picking skills, extendable to other similar tasks.
Objective
The fresh food industry is highly labour-intensive, with labour costs often contributing up to 50% of overall production costs. Pressure is growing to reduce production costs while facing major labour shortages. So far robotic automation for picking of delicate fresh produce has been impossible mainly due to the complex, contact-rich interactions involved in such tasks.
SoftGrip will deliver an innovative soft gripper solution for the autonomous picking of delicate white button mushrooms cultivated on Dutch shelves. The versatility of the proposed solution will enable the adoption of the technology by other fresh-food industries experiencing similar stringent handling requirements such kiwifruit, grapes, etc.
Towards this goal, our consortium will develop: (a) low-cost, soft robotic grippers having built-in actuation, sensing and embodied intelligence that enable reliable and efficient picking of mushrooms; (b) material synthesis and fabrication techniques that offer precise tuning of mechanical properties, comply with food-safe standards, allow for chemical recycling and offer self-repair properties; (c) a set of accelerated continuum mechanics modelling algorithms that facilitate real-time model-based control schemes, capable of being executed by limited computational resources. (d) advanced learning capabilities of the soft gripper through a learning by imitation framework comprising multi-task and meta-learning techniques, so that SoftGrip can be deployed with minimal programming effort.
SoftGrip will enable a step change in efficiency, helping mushroom growers cut down on costs by >30% and increase their yields by >20% while also improving job quality in the industry. In the long-term, it will lower the barriers of robotics deployment open up new opportunities for adoption of robotic solutions in the agri-food sector.
Fields of science
Keywords
Programme(s)
Funding Scheme
RIA - Research and Innovation actionCoordinator
56127 Pisa
Italy