Self-organizing control systems for autonomous educational robotic complexes

Abstract

Abstract The increasing social importance of educational robotics caused development not only a technical component of educational robots. But improvement of control algorithms is more important. At the same time the most important direction of development is the possibility of autonomous functioning at a solution of difficult tasks on dynamically changing external environment. Expansion of orbs of application of automatic control has caused development of intellectualization of control systems. One of the important directions is intelligent self-organizing of automatic control system (ISACS). They are capable to supply required capabilities of the purpose of control with change of environments and/or their parameters. It is attained by automatic synthesis of the control law, the most adequate a current situation. For this purpose the intelligent system of synthesis is used. The planning subsystem creates (in the elementary case selects from already known) the most adequate procedure of synthesis. However, existing approaches to planning actions have no property of mass parallelism. It do not allow applying them in control systems owing to the big costs of time for a solution of task. It is offered to use artificial neural planning networks (ANPN) within the planning subsystem of tasks solver. Features of planning of tasks solving with use ANPN are considered. These abilities become the urgent need for implementation of opportunities of autonomous functioning of socially important educational robots. Outcomes of simulation of control of a plants population with use ISACS are represented.