Thesis/1-goals-and-outcomes/research-statement.tex

% GOAL PARAGRAPH
The goal of this research is to develop a methodology for creating autonomous
\oldt{control systems with event-driven control laws that have guarantees of
safe and correct behavior.} \newt{hybrid control systems with mathematical
guarantees of safe and correct behavior.}

% INTRODUCTORY PARAGRAPH Hook
Nuclear power relies on extensively trained operators who follow detailed
written procedures to manage reactor control. Based on these procedures and
\oldt{operators'} \newt{their} interpretation of plant conditions,
\oldt{operators} \newt{they} make critical decisions about when to switch
between control objectives.
% Gap
\oldt{But, reliance} \newt{This reliance} on human operators has created an
economic challenge for next-generation nuclear power plants. Small modular
reactors face significantly higher per-megawatt staffing costs than
conventional plants.Autonomous
control systems \oldt{are needed that can} \newt{must} safely manage complex
operational sequences with the same assurance as human-operated systems, but
without constant supervision.

% APPROACH PARAGRAPH Solution
To address this need, we will combine formal methods from computer science
with control theory \oldt{to build hybrid control systems that are correct by
construction.} \newt{to build hybrid control systems that are correct by
construction, leveraging the extensive domain knowledge already embedded in
existing operating procedures and safety analyses.}
% Rationale
Hybrid systems use discrete logic to switch between continuous control modes,
similar to how operators change control strategies. Existing formal methods
generate provably correct switching logic but cannot handle continuous
dynamics during transitions, while traditional control theory verifies
continuous behavior but lacks tools for proving discrete switching
correctness.

% Hypothesis and Technical Approach
We will bridge this gap through a three-stage methodology. First, we will
translate written operating procedures into temporal logic specifications
using NASA's Formal Requirements Elicitation Tool (FRET). \oldt{which
structures requirements into scope, condition, component, timing, and
response elements. This structured approach enables realizability checking to
identify conflicts and ambiguities in procedures before implementation.}
\newt{FRET structures requirements into scope, condition, component, timing,
and response elements, enabling realizability checking that identifies
conflicts and ambiguities in procedures before implementation.}

Second, we will synthesize discrete mode switching logic using reactive
synthesis \oldt{to generate deterministic automata that are provably correct
by construction.} \newt{to produce deterministic automata that are correct by
construction.}
Third, we will develop continuous controllers for each discrete mode using
standard control theory and reachability analysis. We will classify
continuous modes based on their transition objectives \oldt{, and then employ
assume-guarantee contracts and barrier certificates to prove that mode
transitions occur safely and as defined by the deterministic automata.}
\newt{and verify safe mode transitions using barrier certificates and
reachability analysis.}

This compositional approach enables local verification of continuous modes
without requiring global trajectory analysis across the entire hybrid system.
\oldt{We will demonstrate this on an Emerson Ovation control system.}
\newt{We will validate this methodology through hardware-in-the-loop testing
on an Emerson Ovation distributed control system, made possible through the
University of Pittsburgh Cyber Energy Center's industry partnership.}

% Pay-off
This approach \oldt{will demonstrate autonomous control can be used for}
\newt{enables autonomous management of} complex nuclear power operations
while maintaining safety guarantees.

% OUTCOMES PARAGRAPHS
If this research is successful, we will be able to do the following:
\begin{enumerate}
  % OUTCOME 1 Title
  \item \textit{Synthesize written procedures into verified control logic.}
    % Strategy
    We will develop a methodology for converting written operating procedures
    into formal specifications. These specifications will be synthesized into
    discrete control logic using reactive synthesis tools.
    % Outcome
    \oldt{Control engineers will be able to generate mode-switching
    controllers from regulatory procedures with little formal methods
    expertise, reducing barriers to high-assurance control systems.}
    \newt{This will reduce barriers to high-assurance control systems by
    generating verified mode-switching controllers directly from regulatory
    procedures.}

    % OUTCOME 2 Title
  \item \textit{Verify continuous control behavior across mode transitions.}
    % Strategy
    We will develop methods using reachability analysis to ensure continuous
    control modes satisfy discrete transition requirements.
    % Outcome
    Engineers will be able to design continuous controllers using standard
    practices while ensuring system correctness and proving mode transitions
    occur safely at the right times.

    % OUTCOME 3  Title
  \item \textit{Demonstrate autonomous reactor startup control with safety
    guarantees.}
    % Strategy
    We will implement this methodology on a small modular reactor simulation
    using industry-standard control hardware.
    % Outcome
    \oldt{Control engineers will be able to achieve autonomy without
    retraining costs or developing new equipment by implementing
    high-assurance autonomous controls on industrial platforms they already
    use.} \newt{Without retraining costs or new equipment, control engineers
    will be able to implement high-assurance autonomous controls on industrial
    platforms they already use.}

\end{enumerate}