/HRI_Tactile

Reinforcement learning on NAO via tactile sensor

Primary LanguagePython

HRI Tactile Project

Overview

A research internship at Institute for Cognitive Systems in TU Munich. With supervisor Mohsen Kaboli. Base on NAO and artificial skin.

  • Discretize the state space of joints and use Q-network to train the robot to do some manipulations
  • DMP policy search to achieve some specific trajectories
  • Demostration