Robotic Manipulation Task Solving Using Cognitively-Inspired Causal Learning
Študent | Miroslav Cibula (cibula25@uniba.sk) |
Školiteľ | prof. Ing. Igor Farkaš, Dr. |
Konzultant | Mgr. Michal Vavrečka, PhD. |
Zadanie
Anotácia
Observing and learning causal relations in a given environment is an essential element
of cognition in humans and high animals. A causal model of the world allows an agent
to predict the effect of its actions on the environment. When combined with mental simulation,
such information can be utilized for creative, flexible, and universal task-solving
in a known environment.
Ciele
-
To design a system enabling a robotic arm (agent) to learn causality by manipulating objects, to represent and store this information, and to apply it in solving tasks in known environment.
-
To implement a simulated randomizable environment for the system training and inference and to test the system on a set of experiments in it.
-
To analyze results from the experiments and the overall system efficiency.
Výstupy
- Text práce (WIP, naposledy aktualizované 1.5.2024)
- Ďalšie výsledky