Linear–Quadratic Regulator

Linear–Quadratic Regulator
Name	Linear–Quadratic Regulator
Field	Control theory
Invented	1960s
Related	Linear–Quadratic-Gaussian regulator, Riccati equation, optimal control

Linear–Quadratic Regulator The Linear–Quadratic Regulator (LQR) is a foundational optimal control design method that computes a state-feedback law minimizing a quadratic cost for linear dynamical systems. It underpins modern control practice across aerospace, robotics, and economics and connects to the algebraic Riccati equation, dynamic programming, and matrix analysis.

Introduction

LQR emerged in the mid-20th century amid advances in systems theory and optimal control associated with institutions like Massachusetts Institute of Technology, Bell Labs, Princeton University, Stanford University, California Institute of Technology, and University of Cambridge. Early contributions came from researchers affiliated with RAND Corporation and programs connected to National Aeronautics and Space Administration and Department of Defense (United States), influencing projects such as Apollo program and control work at Lockheed Martin. LQR is linked historically to figures who worked at places like Bell Telephone Laboratories and Institute for Advanced Study, and to mathematical developments at Courant Institute and Harvard University.

Problem Formulation

The canonical continuous-time LQR problem considers a linear time-invariant plant with state matrices that might be analyzed at facilities such as Brookhaven National Laboratory or taught at École Polytechnique. The cost functional is quadratic in states and controls, paralleling quadratic forms studied by mathematicians at Princeton University and University of Oxford. Constraints and performance specifications often reference standards developed by International Organization for Standardization and defense specifications from NATO, with practical tuning drawing on methods used at General Electric and Siemens.

Solution via Riccati Equation

The optimal state-feedback gain is derived from the solution of the matrix Riccati differential or algebraic equation, a topic with roots in work published by scholars from University of California, Berkeley and Imperial College London. Numerical solvers for Riccati equations are implemented in toolkits maintained by institutions like MathWorks and open-source projects inspired by coding from contributors at Google and Microsoft Research. The theoretical analysis leverages matrix theory advanced by researchers at ETH Zurich and University of Paris.

Properties and Optimality

LQR yields a control law with provable optimality and robustness properties that feature in curricula at Massachusetts Institute of Technology and Stanford University. Stability proofs and performance bounds have been developed in collaboration with groups at National Institute of Standards and Technology and Los Alamos National Laboratory. The closed-loop poles and eigenstructure design principles connect to modal analysis used in companies like Boeing and Airbus.

Extensions and Variants

Extensions such as the Linear–Quadratic–Gaussian (LQG) regulator combine LQR with stochastic estimation techniques from the Kalman filter, an algorithm associated with work at Stanford University and applications in projects like Apollo program. Robust and H-infinity variants build on research from University of California, Santa Barbara and Carnegie Mellon University. Constrained and model predictive control formulations relate to research at Imperial College London and ETH Zurich and have been adopted by firms including Toyota and Siemens.

Applications

LQR is applied widely: attitude control for satellites developed by teams at Jet Propulsion Laboratory and European Space Agency; flight control at Boeing and Airbus; active suspension systems researched at General Motors and Ford Motor Company; robotic manipulators and legged robots designed by labs at Massachusetts Institute of Technology and Carnegie Mellon University; and economic stabilization models studied at Federal Reserve System and International Monetary Fund. Additional uses occur in signal processing research at Bell Labs and Nokia Bell Labs, and power systems control in utilities managed by National Grid (Great Britain) and Pacific Gas and Electric Company.

Numerical Implementation and Examples

Software implementations and numerical examples of LQR appear in environments maintained by MathWorks, integrated into teaching at University of California, Berkeley and showcased in benchmarks by IEEE conferences. Testbeds and demonstration platforms developed at NASA Ames Research Center and Draper Laboratory validate controller performance; industrial case studies originate from Rolls-Royce and General Electric. Computational routines for solving algebraic Riccati equations are distributed in libraries associated with NumPy, contributors from Google, and research groups at University of Illinois at Urbana–Champaign.

Category:Control theory