Correct-by-Construction Autonomy Protocols for Open, Reconfigurable, Networked Naval Systems

Abstract

Abstract: The objective of the proposed effort is to develop methods, computational tools and demonstrations for the formal specification and correct-by-construction synthesis of autonomous control protocols for open, reconfigurable, networked systems. We will utilize electric power distribution and load management systems—as well as their interplay with other physical networks, including fluid flow for thermal management, and information flow networks—on advanced naval platforms and ships for demonstration. The methods and algorithms that we propose to develop for synthesizing autonomous decisionmaking protocols will help enable automation functionalities at higher levels including setting plant configurations; managing configuration transitions; dynamically re-planning as mission requirements and system health conditions change; and optimizing configuration choices based on quantitative tradeoffs. Despite all complicating factors that unsustainably increase the cost of development, the designverification cycle for open, networked systems is often ad hoc and rarely initiated with formal, mathematically-based, unambiguous specifications. It relies on domain expertise and system-level tests (available only at late stages of the design cycle; hence, leading to costly re-design) for establishing assurance. A cause of this primitive state in our capabilities is the lack of computational, (at least partly) automated tools suitable for the design and verification of these systems. Our proposed effort targets this very lack. The tool support we are proposing will help keep the monetary, labor and time costs sustainable in the development and trustworthy deployment of open, reconfigurable, networked systems. The purpose of this seed project is to create a comprehensive plan on the research needs for naval applications. The expected outcomes of this initial phase include the following. - A case study on the joint electric power and fluid network management that includes the salient features of autonomous decision-making for the functionalities discussed above. - Representative formal specifications and performance metrics. - Abstractions that are suitable for the specifications and performance metrics identified for the case study and characterize the interactions with the underlying computing infrastructure and communication resources. - Application of the existing methods (which we had developed in the domain of power distribution on more-electric aerial systems) on the proposed case study. - Identification of a validation plan for a subsequent project in collaboration with and potentially on the testbeds of the Naval Ship Systems Engineering Station.

Document Details

Document Type
DoD Grant Award
Publication Date
Aug 12, 2016
Source ID
N000141512310

Entities

People

  • Ufuk Topcu

Organizations

  • Office of Naval Research
  • United States Navy
  • University of Pennsylvania

Tags

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Cybersecurity.
  • Enterprise Information Systems Architecture and Joint Command Capability Interoperability Support.