Research

Artificially Intelligent 3D-Printed Soft Gripper for Ripeness and Stiffness Identification

Project Details

Abstract

This work reports on a 3D-printed soft robotic gripper equipped with embedded pneumatic sensing chambers designed to handle and categorize tomatoes during the harvesting process. The process of identifying ripeness involves two main stages. First, the gripper makes contact with a tomato using a closed-loop pressure/force control system. This interaction generates a force versus displacement curve, which is fed to a neural network (NN) for the classification of the ripeness of the tomato with an accuracy of 87%. This approach mimics human techniques for assessing the ripeness (i.e., stiffness) of produce, which involves gently touching the produce to identify its ripeness and then handling it based on the identified ripeness level. This soft gripper is ideal for harvesting fruits and vegetables as it can directly identify the ripeness level and apply the required pressure/force to preserve the quality of the produce. In addition, it can be tailored to identify the stiffness of any object depending on the desired application.