ANALYSIS AND SYNTHESIS OF ADVANCED SELF-ORGANIZING CONTROL SYSTEMS - II.
Abstract
This report presents design theory for self-organizing control systems based on the probability state variable (PSV) parameter search algorithm, guided by on-line performance assessment processes. The relationship between the PSV search and general guided random searches is described, and mathematical necessary conditions are derived for self-organization in multiple-input, multiple-actuator controllers. These conditions pertain to controlled plants of arbitrary differential order and are formulated without restriction as to the type and sense of crosscoupling effects within the plant. Synthesis of multiple-input, multiple-actuator self-organizing controller (SOC) networks is discussed and the SOC design procedure is summarized in step-by-step form. Results of application studies are presented. These pertain to simultaneous altitude and attitude control of a VTOL aircraft during hovering flight and to spacecraft simultaneous three-degree-of-freedom orientation control. The simulation data verify that the SOC successfully deals with strongly cross-coupled controlled plants. A number of corollary topics are developed. These include requirements on boundary-value SOC systems; inference of vehicle and atmosphere parameters from free-flight motions; hardware signal differentiation requirements for SOC performance assessment; SOC application to a dual-coil magnetically-torqued satellite (having a constraint on electrical resource consumption); and kinematics of rigid-body rotation, developed using a new quaternions formulation. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Sep 01, 1968
- Accession Number
- AD0840295
Entities
People
- Roger L. Barron