CONVERGE

Convergent Human Learning for Robot Skill Generation

Coordinatore	Ozyegin University    more  Organization address address: NISANTEPE MAH ORMAN SOK 13 city: ALEMDAG CEKMEKOY ISTANBUL postcode: 34794 contact info Titolo: Dr. Nome: Nilay Cognome: Papila Email: send email Telefono: +90 216 564 9568 Fax: +90 216 564 9057
Nazionalità Coordinatore	Turkey [TR]
Totale costo	100˙000 €
EC contributo	100˙000 €
Programma	FP7-PEOPLE Specific programme "People" implementing the Seventh Framework Programme of the European Community for research, technological development and demonstration activities (2007 to 2013)
Code Call	FP7-PEOPLE-2012-CIG
Funding Scheme	MC-CIG
Anno di inizio	2012
Periodo (anno-mese-giorno)	2012-09-03   -   2016-09-02

 Partecipanti

#	participant	country	role	EC contrib. [€]
1	Ozyegin University  Organization address address: NISANTEPE MAH ORMAN SOK 13 city: ALEMDAG CEKMEKOY ISTANBUL postcode: 34794 contact info Titolo: Dr. Nome: Nilay Cognome: Papila Email: send email Telefono: +90 216 564 9568 Fax: +90 216 564 9057	TR (ALEMDAG CEKMEKOY ISTANBUL)	coordinator	100˙000.00

Mappa

 Word cloud

Esplora la "nuvola delle parole (Word Cloud) per avere un'idea di massima del progetto.

 Obiettivo del progetto (Objective)

'The human sensorimotor system enables the learning and control of an ever-growing number of artifacts ranging from simple tools such as chopsticks to advanced computer interfaces. The Researcher’s recent work in Japan exploited this human capacity to obtain dexterous skills on robots, which otherwise would require expert programming. In this framework, a human operator is put in the control loop of a robotic system where (s)he controls the robot in real-time. The operator then ‘learns’ to make the robot perform a given task (e.g. wiping clean a table by controlling a robot). This is analogous to a beginner’s learning to drive a car. After the operator becomes skilled, the signals coming in and leaving out of the robot are used to construct an autonomous controller. The key point of this framework is that it takes away the work from the cognitive system of an expert and places it on a layperson’s sensorimotor system. Up to now, the framework involved a sequential learning scheme: first the human operator learned to control the robot. Then, data was collected when the robot performed the tasked under the skilled operator guidance. Finally, this data was used to learn a policy using a machine learning technique. This project will substantially improve this framework by having the robot simultaneously learn with the human operator. The dynamics of the simultaneous learning of the human operator and the controlled robot will be studied in depth for obtaining quantitative criteria for a convergent learning system. The results of the analysis will be deployed on a robotic platform, and evaluation experiments will be carried out to show that convergent learning can be ensured leading to a virtuous learning experience for both the human and the robot. The development of this framework to its full potential will drastically change how we develop smart prosthetics and build robotic systems that can coexist with humans, for which this project offers an important contribution'