Принятие решений в центральной нервной системе робота

摘要

Introduction: Smart electromechanical systems have the ability to perform parallel calculations, group control, communication, information storage, monitoring, measurement and control of their own or environmental parameters with parallel kinematics of the actuators. This allows such systems to manipulate large loads in terms of accuracy and rigidity. The behavior of such a system is based on the information it obtains from its own central nervous system about the state of the environment and the state of the system itself. Purpose: Developing algorithms of individual behavioral decisions for humanoid robots built from modules of intelligent electromechanical systems. The decisions should be based on the information received from the central nervous system. Results: The paper discusses the deductive, inductive and abductive types of behavioral decision-making in the central nervous system of a robot built from modules of intelligent electromechanical systems. It is shown that the abductive method is the fastest one by analogy with intuition, but its reliability depends on the completeness of the database of good solutions from the past experience, i.e. it strongly depends on the time of operating similar robots in similar environments. The deductive method, with a large number of constraints, is faster than the inductive one, as it does not require that the constraints for all the solutions are verified. Under complex quality criteria and a small number of constraints, the inductive method can produce faster results, as it discards the search for solutions by complex quality criteria for the decisions unacceptable by constraints. Practical relevance: Based on the considered types of behavioral decisionmaking in the central nervous system of a robot, the proposed algorithms can be used to formulate the strategy and tactics of the control over intelligent robots.