Construction by a robot of a logical theory of itself and its environment through experiments

This work studies the possibility of constructing an automatic control system capable of investigating itself and its environment by following a systematic experimentation algorithm and constructing a logical theory to explain the results of these experiments. It is proposed that the behaviour of the world must obey a logical theory. This logical world is equivalent to an “objective” world. On the basis of this second approach, an algorithm is proposed to enable any control system to systematically perform experiments and progressively construct a model of the world. The application of this algorithm is presented step by step in a simple example