Controllability and Observability

Research Depth 111 in the knowledge graph I know this Set as goal
Unlocks 10 downstream topics
controllability observability Kalman-rank PBH-test structural-properties

Core Idea

Controllability determines whether any initial state can be driven to any final state in finite time using the input. The Kalman rank condition states that system (A, B) is controllable if and only if the controllability matrix C = [B AB A²B ⋯ Aⁿ⁻¹B] has full row rank n. Observability determines whether the initial state can be uniquely inferred from the output history; system (A, C) is observable if and only if O = [C; CA; CA²; ⋯; CAⁿ⁻¹] has full column rank n. These properties are dual to each other and can also be tested via the PBH eigenvector test. Controllability is a prerequisite for arbitrary pole placement; observability is required for state estimation.

How It's Best Learned

Construct controllability and observability matrices for 2nd and 3rd order systems and check rank numerically. Practice the PBH test as an alternative verification. Show that changing actuator or sensor location can destroy these properties on the same plant.

Common Misconceptions

Explainer

In state-space representation, the system dynamics are encoded in two objects: the A matrix (how states evolve autonomously) and the B matrix (how inputs influence states). A reasonable assumption might be that since we can pick any input signal, we have complete freedom to push the system anywhere. But this is wrong — certain state variables may be completely hidden from the input, forming "decoupled modes" that evolve independently no matter what we do. Controllability is the formal test for whether this problem exists.

The Kalman controllability matrix 𝒞 = [B, AB, A²B, …, Aⁿ⁻¹B] stacks together all the directions in state space that the input can reach, directly or after 1, 2, up to n−1 steps of propagation through the dynamics. If these columns span all of ℝⁿ — i.e., the matrix has full row rank n — then you can steer the state to any point in finite time using the right input sequence. If the rank is less than n, there exists at least one state-space direction that the input cannot affect, no matter how cleverly you design the control law. The matrix runs up to Aⁿ⁻¹B rather than longer because the Cayley-Hamilton theorem guarantees that Aⁿ and higher powers of A can be expressed as combinations of I, A, …, Aⁿ⁻¹, so no new directions appear beyond n−1 multiplications.

Observability asks the dual question: given the output history y(t) from an unknown initial state, can you uniquely deduce what x(0) was? The observability matrix 𝒪 = [C; CA; CA²; …; CAⁿ⁻¹]ᵀ stacks the output maps after 0, 1, …, n−1 time steps. Full column rank means every distinct initial state produces a distinct output trajectory — they are distinguishable. A rank deficiency means two different initial states produce identical outputs forever, so no observer can tell them apart. The duality is exact: system (A, B) is controllable if and only if (Aᵀ, Bᵀ) is observable. This means any test or design method for one property can be mechanically translated to the other by transposing.

The practical stakes are high. Pole placement — assigning closed-loop eigenvalues via state feedback u = −Kx — requires full controllability. If a mode is uncontrollable, no choice of K can move its eigenvalue; if that mode is also unstable, the system cannot be stabilized by feedback at all. Conversely, building a state observer (estimating x from y) requires full observability: unobservable modes cannot be estimated because they leave no trace in the output. An uncontrollable but stable mode is benign — it decays on its own. An unobservable but stable mode is also manageable — it cannot be estimated but it also doesn't grow. The dangerous pathological case is an uncontrollable unstable mode (cannot be moved by input) or an unobservable unstable mode (grows invisibly in the output without being detected). Good design practice is to verify both properties early, because they depend on actuator and sensor *placement* — not just the dynamics — and a poor placement choice can be impossible to overcome by clever algorithm design.

Practice Questions 5 questions

Prerequisite Chain

Counting to 10Counting to 20Understanding ZeroThe Number ZeroCounting to FiveOne-to-One CorrespondenceCombining Small Groups Within 5Addition Within 10Addition Within 20Two-Digit Addition Without RegroupingTwo-Digit Addition with RegroupingAddition Within 100Repeated Addition as MultiplicationMultiplication Facts Within 100Division as Equal SharingDivision as Grouping (Measurement Division)Division: Grouping (Repeated Subtraction) ModelDivision: Fair Sharing ModelDivision as Equal SharingDivision as GroupingBasic Division FactsDivision Facts Within 100Two-Digit by One-Digit DivisionDivision with RemaindersRemainders and Quotients in DivisionDivision Word ProblemsIntroduction to Long DivisionFactors and MultiplesPrime and Composite NumbersEquivalent FractionsRelating Fractions and DecimalsDecimal Place ValueReading and Writing DecimalsComparing and Ordering DecimalsAdding and Subtracting DecimalsMultiplying DecimalsDividing DecimalsDividing FractionsMixed Number ArithmeticOrder of OperationsInteger Order of OperationsVariable ExpressionsCombining Like TermsOne-Step EquationsTwo-Step EquationsSolving Multi-Step EquationsEquations with Variables on Both SidesAngle Pairs: Complementary, Supplementary, and VerticalParallel Lines and TransversalsCorresponding AnglesAlternate Interior AnglesTriangle Angle Sum TheoremExterior Angle TheoremTriangle Inequality TheoremSimilar Triangles: AA SimilaritySimilar Triangles: SSS and SAS SimilarityProportions in Similar TrianglesRight Triangle Trigonometry IntroductionTrigonometric Ratios ReviewRadian MeasureConverting Between Degrees and RadiansThe Unit CircleGraphing Sine and CosineGraphing Tangent and Reciprocal Trigonometric FunctionsDerivatives of Trigonometric FunctionsAntiderivativesIterated Integrals and Fubini's TheoremDouble Integrals in Cartesian CoordinatesDouble Integrals over Rectangular RegionsDouble Integrals in Polar CoordinatesDouble Integrals: Definition and SetupIterated Integrals and Fubini's TheoremDouble Integrals over Rectangular RegionsDouble Integrals over General RegionsApplications of Double Integrals: Area, Mass, and MomentsTriple Integrals in Cartesian CoordinatesTriple Integrals in Cylindrical and Spherical CoordinatesChange of Variables and the Jacobian DeterminantApplications of Triple Integrals: Volume and MassVector Fields and Their RepresentationsLine Integrals of Vector FieldsGreen's TheoremSurface Integrals and Flux of Vector FieldsSurface Integrals and Flux of Vector FieldsDivergence Theorem: Flux and OutflowDivergence TheoremElectric FluxGauss's LawConductors in Electrostatic EquilibriumCapacitance and CapacitorsDielectricsDielectric Constant and Relative PermittivityElectric Field Inside Dielectric MaterialsDielectric Materials and PolarizationDielectric Susceptibility and PermittivityEnergy Density in Electric FieldsElectric Current and Current DensityElectrical Resistance and ResistivityOhm's Law and Circuit ElementsElectromotive Force (EMF) and BatteriesKirchhoff's Circuit Laws: Voltage and CurrentDC Circuit Network Analysis MethodsTransient Response in RC CircuitsRC CircuitsLC and RLC CircuitsSecond-Order Transient Circuit ResponseFeedback Control FundamentalsLaplace Transform Methods for ControlTransfer Functions and System ModelingState-Space RepresentationState Transition MatrixControllability and Observability

Longest path: 112 steps · 619 total prerequisite topics

Prerequisites (6)

Leads To (4)