Intelligent methods for creating and improving motion primitives for anthropomorphic robots

Most modern methods for controlling the movement of humanoid robots are based on the use of the classical control theory, where control actions on the robot are calculated in real time, taking into account feedback from sensors. However, with an increase in the number of degrees of freedom, the task of controlling in this way becomes difficult. One solution is to use the so-called dynamic motion primitives. Management in this case occurs by selecting the necessary motion primitive from a pre-calculated library of primitives. The purpose of scientific research is to create a mathematical apparatus for the automated construction of libraries of primitive movements of anthropomorphic robots.