Control Theory
| Control Theory | |
|---|---|
| Field | Engineering, Mathematics |
| Subfields | Classical control, Optimal control, Robust control, Nonlinear control |
| Key Concepts | Feedback, Stability, Controllability, Observability |
| Founders | |
| Pioneers | Norbert Wiener, Rudolf Kálmán, Harry Nyquist, Hendrik Wade Bode |
Control theory is an interdisciplinary field of engineering and mathematics that deals with the behavior of dynamical systems with inputs, and how their behavior is modified by feedback. The goal is to control a system so that its output follows a desired reference signal, despite the presence of disturbances and model uncertainty.
History
The origins of control theory trace back to James Watt's centrifugal governor (1788), which used feedback to regulate steam engine speed. In the 19th century, James Clerk Maxwell formally analyzed the stability of governors. The modern era began with Nyquist's stability criterion (1932) and Bode's frequency response analysis (1940s), which gave engineers graphical tools for designing feedback controllers.
The state-space revolution came in the 1960s with Rudolf Kálmán's work on controllability, observability, and the Kalman filter. This shifted the field from frequency-domain to time-domain methods and enabled the treatment of multi-input, multi-output (MIMO) systems. The Linear Quadratic Regulator (LQR) and LQG control emerged from this period.
The 1970s and 1980s saw the rise of robust control (H-infinity, mu-synthesis) and adaptive control. The 1990s brought Model Predictive Control (MPC), which uses an explicit model of the system to optimize future behavior over a horizon — particularly successful in process control and, more recently, robotics.
Classical vs Modern Control
Classical control relies on transfer functions and frequency-domain analysis. A PID controller — proportional, integral, derivative — is the most widely used classical controller, and is found in over 90% of industrial control loops. Its three terms handle present error (P), accumulated past error (I), and predicted future error (D).
Modern control operates in the time domain using state-space representations. A system is described by dx/dt = Ax + Bu, y = Cx + Du, where x is the state vector, u the input, and y the output. This formulation enables powerful techniques like LQR (optimal state feedback), Kalman filtering (optimal state estimation), and MPC (constrained optimal control).
Optimal Control
Optimal control seeks to minimize a cost function over a trajectory. The two main approaches are Dynamic Programming (Bellman's principle of optimality, leading to the Hamilton-Jacobi-Bellman equation) and the Pontryagin Maximum Principle (PMP), which provides necessary conditions for optimality via the Hamiltonian.
The Linear Quadratic Regulator (LQR) is the canonical optimal controller for linear systems with quadratic costs. It provides a closed-form state feedback law u = -Kx where K is computed from the Algebraic Riccati Equation (ARE). LQR guarantees stability and optimality for linear systems.
Model Predictive Control
Model Predictive Control (MPC) solves an online optimization problem at each time step, using a model of the system to predict future behavior over a horizon. Only the first control action is applied, then the horizon shifts and the optimization repeats — a technique called receding horizon control.
MPC's key advantage is its ability to handle constraints on states and inputs explicitly. This makes it ideal for systems where safety limits matter, such as chemical plants, autonomous vehicles, and robotics. Modern variants include Nonlinear MPC (NMPC) for nonlinear dynamics and real-time MPC for embedded systems.
Applications
Control theory is everywhere: from the thermostat in your home to the autopilot in aircraft, from the cruise control in your car to the stabilization of legged robots. In robotics, LQR and MPC are used for walking gaits, manipulation, and trajectory tracking. In aerospace, control theory handles rocket guidance, satellite attitude control, and drone stabilization.
In economics, control theory appears in portfolio optimization and monetary policy. In biology, it models homeostasis and neural feedback loops. The field continues to evolve with reinforcement learning, where learned policies approximate optimal controllers for complex, high-dimensional systems.
See Also
See also
References
- ^ Åström, K. J. & Murray, R. M. (2021). Feedback Systems: An Introduction for Scientists and Engineers. Princeton University Press.
- ^ Kálmán, R. E. (1960). "Contributions to the theory of optimal control." Bol. Soc. Mat. Mexicana, 5: 102–119.
- ^ Rawlings, J. B. & Mayne, D. Q. (2009). Model Predictive Control: Theory and Design. Nob Hill Publishing.