TOOLS FOR RECONFIGURABLE CONTROL SYSTEM COMPARISONS FOR AUTONOMOUS ASSEMBLY APPLICATIONS

摘要

Reconfigurable control systems are a key enabling technology for autonomous assembly, allowing for a single control system to maintain performance despite large mass property variations that occur from docking and undocking during assembly. There is a wide range of reconfigurable control systems that can be implemented, based on the amount of a priori information known to the designer. The range spans from gain scheduling, where most all the information about the transitions are known, to online system identification, where no information is known. However, little to no techniques currently exist to compare these methods to determine which reconfigurable control system design is the appropriate design for the given assembly scenario. This paper offers two tools to enable the comparison which aids in the selection of a control system for assembly: a set of quantitative metrics and an assembly simulation. Results are presented for sixteen reconfigurable control systems methods and tested on two scenarios, a space telescope and spacecraft self-assembly. Results show a preferred control system method to optimize the performance of the assembly, based on the scenario. The tools and results are integrated in a two phase process that guides engineers to select the appropriate design for autonomous assembly.