First published 15 August 2024
QUT researchers, who recognised that grasping objects is difficult for robots, have built a smart object that can measure a robot’s squeezing force.
The QUT Centre for Robotics research team was led by Dr Fangyi Zhang, under the mentorship of Distinguished Emeritus Professor Peter Corke, with contributions from research assistant Maceon Knopke, and former research assistant and QUT masters student Liguo Zhu.
Dr Zhang said robots needed to pick up all sorts of objects, soft and hard, they needed to hold them firmly, not drop them, but also to not damage them by squeezing too hard.
“To find out how hard the robot is squeezing an object we built a smart object that can measure this squeezing force, how hard it is squeezed and where,” Dr Zhang said.
“The paper tackles the difficulty in assessing robotic grasping of ‘compliant objects’ – which are objects that have an ability to yield elastically when a force is applied.
“Things like fruit or other delicate items, where precise force application is critical to prevent slippage or damage.”
Mr Knopke said the research developed a new method to assess such compliant robotic grasping by using special (sensorised) objects that can feel and measure how much force is applied.
“A test design was proposed to use sensors and magnets to measure where and how strongly something is being touched,” Mr Knopke said.
“The design demonstrated promising results, accurately detecting contact locations and applied forces during robotic grasping tasks, highlighting its potential for broader applications for grasping performance evaluation in robotics.
“The research included extensive testing, showcasing the effectiveness of the design in both single and multiple contact scenarios, paving the way for future enhancements,” Mr Knopke said.
Impactful contributions and future implications
Dr Zhang said the work addressed a significant challenge in robotics - measuring stress or damage to objects during handling - by providing direct assessment capabilities through innovative sensor integration.
“The findings open new avenues for developing more sensitive and accurate performance evaluation toolkits, aiding the development and improvement of robotic systems capable of handling delicate objects,” Dr Zhang said.
“Potential applications span various industries including agriculture, healthcare, and manufacturing, where robotic systems must manage fragile items safely.”
Dr Zhang said he and his co-authors provided open access to the supplementary materials, including CAD models and source codes, to help in fostering further research and innovation in the field.
“This work was supported by the QUT Centre for Robotics 2022 ECR and MCR Funding Scheme and the GentleMAN Project funded by the Research Council of Norway, indicating strong international collaboration.”
Main image: From left, (inset) Distinguished Emeritus Professor Peter Corke, Maceon Knopke, and Dr Fangyi Zhang.
Read the paper: Towards Assessing Compliant Robotic Grasping From First-Object Perspective via Instrumented Objects
Media contact:
Pat Whyte, QUT Media, 07 3138 1150, media@qut.edu.au
After hours: 0407 585 901
