Proportional–Integral–Derivative control

Proportional–Integral–Derivative control
Name	Proportional–Integral–Derivative control
Caption	Industrial PID controller panel
Invented	1910s–1940s
Inventor	Multiple
Field	Control engineering
Applications	Process control, robotics, aerospace

Proportional–Integral–Derivative control is a widely used feedback control technique developed and refined during the 20th century for maintaining desired outputs in dynamic systems. It combines three corrective actions—proportional, integral, and derivative—to reduce error between a measured variable and a setpoint, and it has shaped industrial practice across sectors such as chemical processing, power generation, and aerospace. Engineers and researchers from institutions including Massachusetts Institute of Technology, General Electric, Siemens, Honeywell, and California Institute of Technology contributed to its theoretical maturation and industrial adoption.

Introduction

PID control emerged from early work on regulators and governors, influenced by developments at University of Cambridge, Brown Boveri, Bell Laboratories, Imperial College London, and ETH Zurich. Early implementations appeared in mechanical governors studied by James Watt and later in pneumatic and electronic controllers developed by Westinghouse, Allis-Chalmers, and RCA. The technique underpins automation systems designed by firms such as ABB and Schneider Electric and has been standardized in industrial protocols adopted by International Electrotechnical Commission and American National Standards Institute.

Theory and mathematical formulation

The PID algorithm is described by a control law combining proportional, integral, and derivative terms, derived and formalized in texts from Stanford University, Princeton University, University of California, Berkeley, McGill University, and Tokyo University. In continuous time the control signal u(t) = K_p e(t) + K_i ∫ e(t) dt + K_d de(t)/dt, where notation and Laplace-transform methods appear in works from Norbert Wiener's school and lectures at California Institute of Technology and Massachusetts Institute of Technology. Frequency-domain analysis using the Laplace transform, Bode plots popularized by Hendrik Bode at Bell Laboratories, and Nyquist criteria developed by Harry Nyquist at Bell Laboratories provide stability conditions, while pole-placement and root-locus techniques from Walter R. Evans and Rudolf E. Kalman inform controller design decisions taught at Harvard University and Yale University.

Tuning methods

Tuning methods range from empirical heuristic approaches attributed to Ziegler–Nichols (developed with influence from John G. Ziegler and Nathaniel B. Nichols) to optimization-based schemes used in research at IBM Research, Microsoft Research, National Institute of Standards and Technology, and Lawrence Livermore National Laboratory. Ziegler–Nichols closed-loop and open-loop rules, Cohen–Coon formulas derived with contributions from Edward A. Lee-style methodologies, and relay auto-tuning techniques influenced practitioners in companies like Emerson Electric and Rockwell Automation. Modern methods include model-based optimization using algorithms from Richard Hamming's numerical analysis lineage, genetic algorithms inspired by work at Los Alamos National Laboratory, and robust H-infinity tuning associated with Stefan S. Hansen and research groups at Imperial College London.

Implementation and practical considerations

Digital implementations utilize discrete approximations taught in courses at University of Illinois Urbana-Champaign, Cornell University, University of Michigan, and Carnegie Mellon University, using difference equations and anti-windup schemes discussed in literature from Siemens and Honeywell. Practical issues addressed by engineers at General Motors and Boeing include sampling rate selection, sensor noise mitigation using filters influenced by Andrey Kolmogorov-era stochastic theory, actuator saturation handling seen in NASA flight control practices, and embedded deployment on microcontrollers from Intel, ARM Holdings, and Texas Instruments. Safety and certification considerations reference standards promulgated by Federal Aviation Administration and European Union Aviation Safety Agency for aerospace and by International Organization for Standardization for industrial systems.

Variations and extensions

Extensions include PI, PD, and lead–lag compensators popularized in curricula at University of Oxford and University of Cambridge, as well as adaptive and auto-tuning PID variants developed in labs at Stanford University, MIT Lincoln Laboratory, Draper Laboratory, and ETH Zurich. Fractional-order controllers drawing on research at Karlsruhe Institute of Technology and sliding-mode strategies influenced by Vladimir Utkin complement robust control frameworks advanced by John C. Doyle and Gunter Stein. Cascaded PID architectures used by Shell and ExxonMobil in process control and feedforward-plus-feedback designs employed at Lockheed Martin illustrate practical hybridizations.

Applications

PID controllers regulate temperature in systems designed by Siemens and Honeywell, speed in drives by Mitsubishi Electric and Siemens, and attitude in spacecraft developed by European Space Agency and NASA. They are integral to automotive systems produced by Toyota, Ford Motor Company, and General Motors, to robotic manipulators from KUKA and ABB Robotics, and to semiconductor fabrication equipment by ASML and Applied Materials. Research applications appear in experiments at CERN, Max Planck Society, Lawrence Berkeley National Laboratory, and Johns Hopkins University.

Limitations and stability analysis

Limitations include degraded performance with nonlinearity, time delay, and high-frequency noise—challenges studied by researchers at Princeton University and Caltech. Stability margins are assessed via Nyquist and Bode methods formulated at Bell Laboratories and expanded in robust control theory from UC Berkeley and MIT CSAIL. Anti-windup, bumpless transfer, and gain scheduling used by Rolls-Royce and Raytheon Technologies mitigate actuator constraints and operating-point variations; when those measures are insufficient, advanced methods from Delft University of Technology and ETH Zurich such as model predictive control are preferred.

Category:Control theory