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

\section{Goals and Outcomes}

% GOAL PARAGRAPH
The goal of this research is to develop a methodology for creating autonomous
hybrid control systems\footnote{A \textit{hybrid control system} combines two
types of control: discrete decisions (like ``switch from heating mode to
cooling mode'') and continuous control (like gradually adjusting a
temperature). Most complex systems---cars, aircraft, power plants---work this
way, switching between different operating modes while smoothly controlling
physical processes within each mode.} with mathematical guarantees of safe and
correct behavior.

% INTRODUCTORY PARAGRAPH Hook
Nuclear power plants require the highest levels of control system reliability,
where failures can result in significant economic losses, service interruptions,
or radiological release.
% Known information
Currently, nuclear plant operations rely on extensively trained human operators
who follow detailed written procedures and strict regulatory requirements to
manage reactor control. These operators make critical decisions about when to
switch between different control modes based on their interpretation of plant
conditions and procedural guidance.
% Gap
This reliance on human operators prevents autonomous control and creates a
fundamental economic barrier for next-generation reactor designs. Small modular
reactors\footnote{\textit{Small modular reactors} (SMRs) are a new generation
of nuclear reactors that are physically smaller than traditional plants and can
be factory-built in modules. Think of the difference between building a custom
house on-site versus assembling a prefabricated one. They produce less power
individually but are designed to be cheaper and faster to deploy.} face
per-megawatt staffing costs far exceeding those of conventional plants,
threatening their economic viability.

% Critical Need
What is needed is a method to create autonomous control systems that safely
manage complex operational sequences with the same assurance as human-operated
systems, but without constant human supervision.
% APPROACH PARAGRAPH Solution
To address this need, we will combine formal methods\footnote{\textit{Formal
methods} are mathematical techniques used to prove that a system will behave
exactly as intended---not just test it and hope, but actually \textit{prove}
it the way you prove a theorem in geometry. If the proof holds, the system
cannot have certain types of errors. This is the gold standard for
safety-critical systems.} with control theory to build hybrid control systems
that are correct by construction.\footnote{\textit{Correct by construction}
means the system is built in a way that guarantees correctness from the start,
rather than building something and then testing to find bugs. The design
process itself prevents errors from being introduced.}
% Rationale
Hybrid systems use discrete logic\footnote{\textit{Discrete logic} deals with
distinct, separate states---like an on/off switch or a set of step-by-step
instructions. This is in contrast to \textit{continuous} behavior, which
changes smoothly over time, like temperature rising gradually. The challenge
Dane is tackling is that nuclear reactors involve \textit{both}: operators
follow step-by-step procedures (discrete) that control smoothly changing
physical processes (continuous).} to switch between continuous control modes,
mirroring how operators change control strategies. Existing formal methods can
generate provably correct switching logic from written requirements, but they
cannot handle the continuous dynamics that occur during transitions between
modes. Meanwhile, traditional control theory can verify continuous behavior but
lacks tools for proving correctness of discrete switching decisions.
% Hypothesis
By synthesizing discrete mode transitions directly from written operating
procedures and verifying continuous behavior between transitions, we can create
hybrid control systems with end-to-end correctness guarantees. If existing
procedures can be formalized into logical specifications and continuous dynamics
verified against transition requirements, then autonomous controllers can be
built that are provably free from design defects.
% Pay-off
This approach will enable autonomous control in nuclear power plants while
maintaining the high safety standards required by the industry. The University
of Pittsburgh Cyber Energy Center's partnership with Emerson provides access to
industry-standard control hardware, ensuring that developed solutions align with
practical implementation requirements from the outset.

% OUTCOMES PARAGRAPHS
If this research is successful, we will be able to do the following:

\begin{enumerate}

  % OUTCOME 1 Title
  \item \textbf{Translate written procedures into verified control logic.}
    % Strategy
    We will develop a methodology for converting existing written operating
    procedures into formal specifications\footnote{A \textit{formal
    specification} is a precise, mathematical description of what a system
    must do. Written operating procedures say things like ``if temperature
    exceeds 315\textdegree{}C, switch to cooling mode.'' A formal specification
    says the same thing in mathematical language that a computer can reason
    about and verify.} that can be automatically synthesized
    into discrete control logic. This process will use structured intermediate
    representations to bridge natural language procedures and mathematical
    logic.
    % Outcome
    Control system engineers will generate verified mode-switching controllers
    directly from regulatory procedures, lowering the barrier to high-assurance
    control systems.

    % OUTCOME 2 Title
  \item \textbf{Verify continuous control behavior across mode transitions.}
    % Strategy
    We will establish methods for analyzing continuous control modes to ensure
    they satisfy discrete transition requirements. Using classical control
    theory for linear systems and reachability analysis\footnote{\textit{Reachability analysis} answers the question: ``Starting from
    here, what are all the possible places the system could end up?'' If you
    can show that all possible paths stay within safe boundaries and eventually
    reach the target, you have proven the controller works correctly.} for
    nonlinear dynamics, we will verify that each continuous mode safely reaches
    its intended transitions.
    % Outcome
    Engineers will design continuous controllers using standard practices
    while iterating to ensure broader system correctness, proving that mode
    transitions occur safely and at the correct times.

    % OUTCOME 3  Title
  \item \textbf{Demonstrate autonomous reactor startup control with safety
    guarantees.}
    % Strategy
    We will apply this methodology to develop an autonomous controller for
    nuclear reactor startup procedures, implementing it on a small modular
    reactor simulation using industry-standard control hardware. This
    demonstration will prove correctness across multiple coordinated control
    modes from cold shutdown through criticality\footnote{\textit{Criticality}
    is the point at which a nuclear reactor sustains a chain reaction on its
    own. Getting there safely from a cold, shut-down state involves carefully
    coordinated steps---this is the startup sequence Dane aims to automate.}
    to power operation.
    % Outcome
    We will demonstrate that autonomous hybrid control can be realized in the
    nuclear industry with current equipment, establishing a path toward
    reduced operator staffing while maintaining safety.

\end{enumerate}

% IMPACT PARAGRAPH Innovation
The innovation in this work is unifying discrete synthesis with continuous
verification to enable end-to-end correctness guarantees for hybrid systems.
% Outcome Impact
If successful, control engineers will create autonomous controllers from
existing procedures with mathematical proof of correct behavior. High-assurance
autonomous control will become practical for safety-critical applications.
% Impact/Pay-off
This research will provide the tools to achieve that autonomy while maintaining
the exceptional safety record the nuclear industry requires.