GARTHIM. Remote Reinforcement Learning for Mobile Robots Connected to Internet

 

This project addresses the execution of the cognitive processes of learning and decision making in mobile robots with low on-board computational power, in particular by running the processes in remote stations.

 

This project addresses the execution of the cognitive processes of learning and decision making in mobile robots with low on-board computational power, in particular by running the processes in remote stations.

Currently, learning to make decisions is a very difficult task for a mobile robot due to its strong interaction with the real world and the computational and physical costs of existing methods. In this project we will research in the implementation of those costly processes in a remote station to which the robot is connected through the Internet, and in the implications of the stochastic delays induced by that configuration in learning and decision making.

The project officially begins in November the 15th, 2019, and lasts for 2 years

REFERENCE: UMA18-FEDERJA-113
TITLE: GARTHIM. REMOTE REINFORCEMENT LEARNING FOR MOBILE ROBOTS CONNECTED TO INTERNET
PRINCIPAL RESEARCHER: JUAN-ANTONIO FERNÁNDEZ-MADRIGAL
INSTITUTION: UNIVERSITY OF MÁLAGA
DEPARTMENT: SYSTEM ENGINEERING AND AUTOMATION
DATES: 15/11/19 to 15/11/21
TEAM:

  • Dr. Juan-Antonio Fernández-Madrigal (PI)
  • Dr. Alfonso García-Cerezo
  • Dr. Ana Cruz-Martín
  • Dr. Vicente Arévalo-Espejo
  • Manuel Castellano Quero (PhD Student)

Institutional information (in Spanish): Proyecto de Investigación en el marco del operativo FEDER Andalucía 2014-2020, convocatoria 2018, código UMA18-FEDERJA-113; evaluado con 80 sobre 100; concedido por Resolución definitiva de 10 de octubre de 2019 del Rector de la Universidad de Málaga; con dotación de 63.353,98 € a ejecutar en 2 años.

Publications

Publications in the scope of this project: