Jeremy Straub

Date of Award

January 2016

Document Type


Degree Name

Doctor of Philosophy (PhD)


Computer Science

First Advisor

Ronald Marsh


This dissertation makes two contributions to the use of the Blackboard Architecture for command. The use of boundary nodes for data abstraction is introduced and the use of a solver-based blackboard system with pruning is proposed. It also makes contributions advancing the engineering design process in the area of command system selection for heterogeneous robotic systems. It presents and analyzes data informing decision making between centralized and distributed command systems and also characterizes the efficacy of pruning across different experimental scenarios, demonstrating when it is effective or not. Finally, it demonstrates the operations of the system, raising the technology readiness level (TRL) of the technology towards a level suitable for actual mission use.

The context for this work is a multi-tier mission architecture, based on prior work by Fink on a “tier scalable” architecture. This work took a top-down approach where the superior tiers (in terms of scope of visibility) send specific commands to craft in lower tiers. While benefitting from the use of a large centralized processing center, this approach is limited in responding to failures and interference.

The work presented herein has involved developing and comparatively characterizing centralized and decentralized (where superior nodes provide information and goals to the lower-level craft, but decisions are made locally) Blackboard Architecture based command systems. Blackboard Architecture advancements (a solver, pruning, boundary nodes) have been made and tested under multiple experimental conditions.