This dissertation explores the development and evaluation of mechanical grippers with flexure-driven eversion mechanism fingers, designed to address the challenges of grasping objects in confined spaces—a common issue in the agri-food and food processing industry that hinders rob
...
This dissertation explores the development and evaluation of mechanical grippers with flexure-driven eversion mechanism fingers, designed to address the challenges of grasping objects in confined spaces—a common issue in the agri-food and food processing industry that hinders robotic automation. Conventional mechanical grippers face significant difficulties in dense environments, such as within piles or plants, as their grasping process relies on positioning fingers around the object performing inward-outward movements from its sides. This approach requires considerable clearance, often obstructed by neighboring obstacles. To overcome this challenge, this research introduces in Chapter 1 a novel grasping approach in which the gripper fingers move tangentially along the surface of the object, utilized as flexure-driven eversion mechanisms, to set or release the grasp while reducing environmental disturbances. This approach significantly reduces the required space, as well as unwanted displacements and forces in the surroundings, and leverages environmental interactions to aid finger navigation.
The main objective is to develop and evaluate flexure-driven eversion mechanism fingers for mechanical grippers, aimed at grasping food in confined spaces, as presented in Chapter 1. The finger design incorporates a curved flexure as the structural backbone, which navigates along the object’s surface—displacing adjacent obstacles if necessary—and securely holds it once enclosed. The flexure is combined with the principles of eversion mechanisms to reduce friction forces during operation by ensuring zero relative speed differences with the environment.
This thesis is divided into two parts. The first part, presented in Chapter 2, investigates the applicability of flexures within the proposed finger concept by analyzing their ability to navigate predictably through multiple obstacles when driven from the base. A case study was conducted, modeling the kinematics and kinetics of a straight base-driven flexure, which bends and moves through two consecutive circular obstacles when pushed forward, using a pseudo-rigid-body modeling (PRBM) approach. This analysis required an extension of existing PRBM techniques, enabling the analysis of flexures in event-based scenarios (e.g., abrupt load changes) and with multiple loads at unknown locations along their length. This was achieved by systematically switching and stitching PRBM-topologies while maintaining the system’s potential energy within reasonable bounds. Experimental validation confirmed the compatibility of base-driven flexures with the proposed finger concept, demonstrating their ability to navigate through obstacles in a predictable manner, with both trajectory and interaction forces modeled with high accuracy for design purposes.
In the second part of this thesis, two finger designs were developed and manufactured as prototype grippers, which were experimentally validated based on defined performance criteria. Chapter 3 presents a preliminary gripper design with three fingers, each implemented as a Dual-Belt Curved-Flexure Eversion Mechanism (DBCF-EM)—a single curved base-driven flexure, with its inner and outer contours covered by two everting (outwardly unrolling) belts guided through rollers. Chapter 4 presents an evolved gripper design featuring Sleeved Concentric-Flexure Eversion Mechanisms (SCF-EM) as fingers—a base-driven channeled backbone comprising two concentric curved flexures, fully enclosed along its circumference by a latex sleeve with engineered stretchability, which everts from the channel. The design is completed by a cable-pulley system that synchronizes both movements.
The results showed that both grippers operated as intended, following the object’s surface and navigating adjacent obstacles, and successfully validated the proposed grasping method, enabling form-closed grasps of densely packed objects. Practical tests involving integration into a robotic set-up tasked with emptying a crate of tomatoes, one at a time, demonstrated an exceptionally high grasp success rate, previously unobserved. While the gripper with DBCF-EM fingers occasionally encountered issues such as mis-grasps, object damage, and hygiene concerns due to its open structure, the gripper with SCF-EM fingers achieved a 100% success rate in picking and placing tomatoes without damaging them or the surroundings. Additionally, the SCF-EM fingers offered improved hygiene compliance, durability, and robustness due to the enclosed eversion sleeve creating a barrier and a reduction in the number of mechanical components. Technical tests of both designs measured relatively low disturbance forces and sufficient holding forces. Comparative studies showed that the integration of the eversion mechanism reduced friction forces by around 90%, reducing both damage and mis-grasps.
In general, these findings demonstrate the effectiveness of the proposed finger designs for grasping objects in confined spaces and their suitability to gently and securely handle fragile agri-food products while accommodating natural variations, although additional engineering, testing, and refinement are required to ensure industrial readiness. Ultimately, the most important outcome is that this dissertation introduces flexure-driven eversion mechanisms for the first time and convincingly demonstrates their exceptional potential for navigation in complex and confined environments. This breakthrough constitutes the central scientific contribution of this work, arising from the need for alternative gripper methods for grasping tasks in tight spaces. This contribution provides a solid foundation for further applications in domains such as medical devices, search-and-rescue robotics, and inspection systems for hard-to-reach environments.