Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Engage Grants Program

University of Waterloo

$25,000.00

Fast contact models to support grasp learning.

EF 2016-17

Résumé
Advanced robotics is an area of rapid technological development and economic growth. A major current challenge is enabling robots to decide on their own how to grasp unfamiliar objects (for example, which part of the object to grasp and how to orient the gripper). This is quite different from the capabilities of traditional assembly-line robots, which perform repetitive movements that are specified in detail by humans. Autonomous grasping capability is essential to many uses of robots outside an assembly line environment. One way to help robots learn how to grasp is to perform millions of computer simulations of grasps, and use machine learning methods to generalize from these simulations to predict which grasps will work for new objects. However, grasp simulations have so far either been too slow or not accurate enough, particularly with respect to simulation of the friction forces between object and gripper. A new type of fast and accurate friction model hasbeen developed recently, but so far it has only been used with simple shapes (such as spheres), not with the wide variety of complex shapes that a robot may have to grasp. This project will investigate whether this new approach can be adapted to complex shapes. We will first collect extensive data on friction forces between a robot gripper and a variety of objects, using a high-throughput automated setup. We will use this data to fit parameters of a sophisticated friction model. We will then combine this model with methods for integrating over complex shapes from computer graphics, in order to predict friction forces for fast grasp simulations, and we will evaluate the accuracy of these simulations. If this approach is successful, it will provide grasp simulations of unparalleled speed and accuracy, facilitating the development of world-leading autonomousgrasping capabilities.
Programme
Engage Grants Program
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data

Tel que publié

Application id
494586-2016
Name nom
Tripp, Bryan
Department département
Systems Design Engineering
Organization id
33
Institution établissement
University of Waterloo
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2016
Competition year année de concours
2016
Award amount
25000
Program id
EGP
Program name en
Engage Grants Program
Program name fr
Programme de subventions d'engagement partenariat
Group en
Research partnerships programs
Group fr
Programmes de partenariat
Committee code
1553
Committee name en
Ontario Internal Decision Committee
Committee name fr
Comité de décision interne pour l'Ontario
Area of application code
802
Area of application group en
Information and communication services
Area of application group fr
Services d'information et de télécommunications
Area of application en
Information systems and technology
Area of application fr
Systèmes d'information et technologie de l'information
Research subject code
2605
Research subject group en
Robotics
Research subject group fr
Robotique
Research subject en
Pattern analysis and machine intelligence
Research subject fr
Analyse de formes et intelligence artificielle
Partie
1
Nb partie
1
Application title
Fast contact models to support grasp learning.
Keyword
NSERC
Application summary
Advanced robotics is an area of rapid technological development and economic growth. A major current challenge is enabling robots to decide on their own how to grasp unfamiliar objects (for example, which part of the object to grasp and how to orient the gripper). This is quite different from the capabilities of traditional assembly-line robots, which perform repetitive movements that are specified in detail by humans. Autonomous grasping capability is essential to many uses of robots outside an assembly line environment. One way to help robots learn how to grasp is to perform millions of computer simulations of grasps, and use machine learning methods to generalize from these simulations to predict which grasps will work for new objects. However, grasp simulations have so far either been too slow or not accurate enough, particularly with respect to simulation of the friction forces between object and gripper. A new type of fast and accurate friction model hasbeen developed recently, but so far it has only been used with simple shapes (such as spheres), not with the wide variety of complex shapes that a robot may have to grasp. This project will investigate whether this new approach can be adapted to complex shapes. We will first collect extensive data on friction forces between a robot gripper and a variety of objects, using a high-throughput automated setup. We will use this data to fit parameters of a sophisticated friction model. We will then combine this model with methods for integrating over complex shapes from computer graphics, in order to predict friction forces for fast grasp simulations, and we will evaluate the accuracy of these simulations. If this approach is successful, it will provide grasp simulations of unparalleled speed and accuracy, facilitating the development of world-leading autonomousgrasping capabilities.