Force-controlled robotic gripping using real-time friction, force, and torque feedback

Download files
Access & Terms of Use
open access
Embargoed until 2019-12-01
Copyright: Wen, Han
Altmetric
Abstract
As robotic applications move towards unstructured environments, there is a need for grippers that can manipulate objects securely (preventing slip) and efficiently (applying minimal grip forces). One approach for achieving this involves measuring the coefficient of static friction (COE) at the gripper-object interface. In the absence of torque, the minimum grip force to prevent slip can be estimated from COE and the contact forces . However, torque at the gripper-object interface is generated when the lifting force does not align with the weight vector of the object being lifted. Current robotic gripping systems ignore the need to measure COE, and there is little literature on the grip force required to prevent slip in the presence of a tangential torque (T) at the gripping interface. In this thesis, a grip force control system was developed, using the measurement of COE at the first contact between the gripper and the object, and the continuous measurement of contact forces and torque to approach the minimal grip force required to hold the object stably. The target grip force is determined in real-time during object manipulation. To demonstrate the importance of COE when no torque is present, objects were gripped with the target grip force dependent only on the measured loads and COE. Furthermore, a model was developed in which the minimum grip force preventing slip can be estimated based on COE, load force, and T. This friction model is first validated with respect to the grip force at which slip is predicted to occur, for varying COE, load force, and T. Objects were then gripped (varying load, COE, and T) with the target grip force calculated by the friction model. The results demonstrate that COE-dependent grip force control is superior to gripping without knowledge of COE, irrespective of T at the contact interface, with respect to the system’s ability to prevent slip and simultaneously minimise the grip force applied. The minimum grip force estimation model and grip force controller developed and validated in this thesis have highlighted the value of measuring COE and the necessity of countering torque to ensure a secure and efficient grip.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Wen, Han
Supervisor(s)
Khamis, Heba
Redmond, Stephen
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2018
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download public version.pdf 12.22 MB Adobe Portable Document Format
Related dataset(s)