This project concerns artificial intelligence in autonomous systems such as robots, and more precisely reinforcement learning.
Nowadays we aim to do ever more intelligent and autonomous robots. One good source of inspiration are the living beings on this planet; some simple isolated systems that once brought together form a strong and reliable system. Trying to adapt it to our robots is a real challenge. And we, humans programming them, can’t think about everything the system may encounter during his life cycle. That is why we try to make the system take decisions based on other criteria such as its past experience; we make the robot learn on its own. But at some point the knowledge acquired depends on the environment, so, what if we modify the system’s environment, how could the robot respond to it quickly? Here, starting from reinforcement learning to rate the decisions, and using adaptive learning algorithms for the gain and loss reward; we aim to make the system responds to the changing environment and adapt its knowledge as time passes.