Agir, Anticiper, s'Adapter

The building of autonomous robot controller in open environments raises the challenging issue of providing the robot with a world model. The approach investigated in Robea AAA is based on Machine Learning and Anticipation: an operational model of the world is obtained by predicting the sensor values
at the next time step, based on the current sensor values and the motor actions.

Activités de recherche
  Apprentissage
  Evolutionary computation

Membres LRI