Use this URL to cite or link to this record in EThOS:
Title: On the science of grasping : modelling grasp affordances in robotics from human analysis
Author: Cotugno, Giuseppe
ISNI:       0000 0004 7427 9059
Awarding Body: King's College London
Current Institution: King's College London (University of London)
Date of Award: 2018
Availability of Full Text:
Access from EThOS:
Access from Institution:
This thesis investigates the human motion patterns of grasping and approaching, as well as applies the functional features to a robotic grasp a ̇ordances controller. Human grasping has always been seen as a target milestone in robotics; however, it is still di°cult for robots to perform advanced grasping tasks, such as required in an assembly line. For robots to operate in an unstructured environment, it is necessary to focus the attention on performing a speci ̋c action with any available object by grasping it in the most pro ̋cient way. The neuroscienti ̋c application of the theory of a ̇ordances explains how this process happens for humans. This thesis analyses human behaviours of grasping and approaching to grasp when performing an action. This study is needed to understand the common behavioural and morphological factors that in ̨uence the motion planning and selection of grasp postures and grasp a ̇ordances. The main features of motion are statistically analysed and adapted to robotic application. A human inspired approaching and grasping controller is proposed, and the limitations and bene ̋ts of state of the art robotic hands are derived based on human data. The approach of this thesis di ̇ers from the traditional method of studying grasp a ̇ordances. Di ̇erent elements of the problem, such as perception, learning, mo-tor control and intention, are modularly separated in components rather than monolithically aggregated in a single entity. A biologically motivated modular grasp a ̇ordance system can adapt only selected features of human grasping to robotics, leaving out biological elements not needed in robotic applications. The modularity of the system separates cognitive motor decisions from the robot's own embodiment and the geometric properties of the objects, granting more in-dependence from the speci ̋c domain of application.
Supervisor: Sklar, Elizabeth Ida ; Nanayakkara, Thrishantha Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available