ANALYSIS AND SYNTHESIS OF ADVANCED SELF-ORGANIZING CONTROL SYSTEMS.
Final rept. 1 Feb 66-28 Feb 67,
ADAPTRONICS INC MCLEAN VA
Pagination or Media Count:
This report presents a major part of the theory of multiple-goal, multiple-actuator self-organizing control systems and results of simulations for several representative applications. The approach taken by the authors is to view self-organizing control as a process where, in general, multiple actuators are driven with respect to multiple performance goals without explicit a priori knowledge as to the functional and quantitative relationships between the goals and the effects of the actuators on them. Probability state variable PSV parameter space search techniques are employed in Identification Logic of the self-organizing controller SOC. The high-speed random search realized with PSV techniques is instrumental in accomplishing simultaneous on-line identification of multiple parameters of the controlled plant. Topics in SOC network connectivity, dynamic coupling between controlled variables, nonlinear coordinate transformations, performance assessment, parameter search, statistical correlation, and SOC modeling and simulation are considered. Spacecraft three-axis attitude control is the principal application treated in this report. It is shown that the SOC controls simultaneously all three axes of a spacecraft without fundamental constraints on the magnitude of angular rates or initial attitude errors.