Abstract:
This article proposes a project that integrates models of a virtual physically correct environment and a humanoid agent for the purpose of synthesis, capable of learning intelligent, including goal-directed behavior. The work formulates the requirements for such a system, in particular the requirements for modeling spatial geometry, pseudo-real time, the dynamics of a multimodal representation of matter for a virtual environment and the requirements for the complex structure of the agent and its control system, the physical correctness of modeling its body and actions in the environment, imitation of life activity, goal setting and genetic evolution.