Software, 2023, ENG

Robot personalities adaptation system

Zedda E.

CNR-ISTI, Pisa, Italy

Description: An application with an adaptation technique implemented on a humanoid robot to make it autonomously adapt its behaviour during user interaction. The application read the user's state (smile value and gaze direction), classified the engagement level following a defined classification and adapted its behaviours using a Q-learning algorithm (Reinforcement Learning algorithm) performing different robot movements and dialogues according to the user state detected and robot personality chosen. The adaption system is applied at a difficult level of a cooking serious game. Technologies used: Java, QiSDK, SQL, Python CNR Link: https://gitea-s2i2s.isti.cnr.it/Human_in_Information_Systems_HIIS/MediumAdaptation

Keywords

Robot behaviuor adaption, Q-learning, Reinforcement learning, Adaptation robot personality

CNR authors

Zedda Eleonora

CNR institutes

ISTI – Istituto di scienza e tecnologie dell'informazione "Alessandro Faedo"