Kushner–Stratonovich Equation¶

Evolve a hidden continuous-time state's normalized conditional law by combining generator-driven prediction with an observation-filtration innovation correction weighted by conditional covariance.

Version: v3 · 2026-09-19 · History
Domain-specific #: 2146
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Nonlinear Stochastic Filtering → Mathematics
Aliases: Kushner equation, Stratonovich–Kushner equation

Core Idea¶

The Kushner–Stratonovich equation is the exact continuous-time nonlinear filtering equation for the normalized conditional law of a hidden stochastic state given its noisy observation history. In a basic independent-noise model,

\[ dX_t=b(X_t,t)\,dt+\Sigma(X_t,t)\,dW_t, \qquad dY_t=h(X_t,t)\,dt+D_t\,dV_t, \]

where W and V are independent Wiener processes and R_t=D_tD_t^T is nonsingular. For a test function varphi, define

\[ \pi_t(\varphi)=E[\varphi(X_t)\mid\mathcal F_t^Y], \]

where F_t^Y contains observations only through time t. If L_t is the backward generator of the hidden signal, then, under the standard hypotheses that the signal and observation SDEs are well posed and nonexplosive, their coefficients satisfy the needed measurability, growth, and integrability conditions, R_t is positive definite, the initial state is independent of the driving Wiener processes, W and V are independent, and varphi lies in the generator domain with integrable terms,

How would you explain it like I'm…

The Guess-Cloud Updater

Imagine a kitten hiding somewhere in a room, and you can only hear fuzzy little sounds. You keep a 'maybe map' of where the kitten could be. As time goes on, you spread your guess out because kittens wander, and each time you hear a sound, you move your guess toward where it fits best. The Kushner–Stratonovich Equation is the exact math rule for updating that maybe map all the time.

Predict-Then-Correct Belief Rule

Sometimes we want to track something we can't see directly, using measurements that are noisy. Instead of one guess, we keep a whole spread of possibilities with how likely each is. The Kushner–Stratonovich Equation says how that spread changes moment by moment. One part predicts: it moves the possibilities the way the hidden thing tends to move by itself. The other part corrects: it compares the new measurement to what we expected and shifts the spread more when our guesses disagree about what we'd measure. The total chance always stays at 100 percent.

Exact Nonlinear Filtering Equation

The Kushner–Stratonovich Equation is the exact continuous-time rule for nonlinear filtering: tracking a hidden random process from a noisy stream of observations. The hidden state follows a random differential equation, and the observation adds noise to some function h of the state. The equation updates the conditional probability distribution of the state given all observations so far. Its first term is prediction, moving the distribution according to the state's own dynamics. Its second is correction: the 'innovation', the observed change minus the expected change, is multiplied by a gain built from the posterior covariance between the quantity of interest and h, scaled by the inverse observation-noise covariance. Unlike a single best guess, it evolves the whole distribution, it is nonlinear in that distribution, and it keeps total probability equal to one.

Consider a hidden diffusion dX = b dt + Σ dW observed through dY = h(X) dt + D dV with independent Wiener processes and nonsingular R = DDᵀ. The Kushner–Stratonovich equation governs π_t(φ) = E[φ(X_t) | observations up to t]. Its weak form is dπ_t(φ) = π_t(Lφ) dt + (π_t(φhᵀ) − π_t(φ)π_t(hᵀ)) R⁻¹ (dY − π_t(h) dt), where L is the signal's generator. The first term is prediction, propagating the law under the dynamics; the second is correction, driven by the innovation dI = dY − π_t(h) dt and weighted by a conditional covariance gain. The innovation has quadratic variation R dt, and whitening it by R^(-1/2) gives a Wiener process in the observation filtration. The equation is nonlinear in π because of products like π(φ)π(h), yet setting φ = 1 shows normalization is preserved. It is best stated for measures, since a density may not exist; when it does, it satisfies dp = L*p dt + p(h − ĥ)ᵀR⁻¹(dY − ĥ dt).

Scope of Application¶

Nonlinear diffusion filtering. The canonical setting is a hidden diffusion observed continuously in additive Gaussian noise. The equation recursively updates every conditional expectation, or the whole density when one exists.

Control under partial observation. A controller can depend only on the observation filtration. The filter converts raw sensor history into a belief state on which separated control policies can operate, subject to the relevant separation theorem and regularity conditions.

Clarity¶

Three distinctions prevent most mistakes.

First, L and L* act on different objects. In weak form, L acts on a test function before expectation. In density form, the forward adjoint L* acts on the density. Writing the backward generator directly on p_t silently changes the equation.

Manages Complexity¶

The raw inference problem asks for the full conditional distribution of every possible hidden path given an ever-growing sensor history. The equation compresses that history into a recursive sufficient state: the current conditional law. Prediction and correction then reuse the same two operators at every instant rather than recomputing Bayes' rule over the entire path.

Abstract Reasoning¶

Use this protocol:

Specify the hidden signal SDE and its generator.
Specify the observation SDE, noise covariance, and correlation assumptions.
Declare the observation filtration and normalized conditional law.
Choose weak test-function form unless density existence is established.
Compute the predicted observation and conditional covariance gain.
Form the innovation using only information available through time t.
Verify normalization with the constant test function.
Test whether a chosen finite-dimensional family is preserved before closing the filter on moments or parameters.

Knowledge Transfer¶

Literal transfer occurs across nonlinear diffusion estimation, target tracking, partially observed control, and continuous-time signal processing when the hidden signal, observation filtration, normalized conditional law, innovation, and covariance gain remain the same mathematical roles. Different applications change b, Sigma, h, and R, not the filtering grammar.

Transfer to Kalman–Bucy is specialization: the posterior family closes under the same equation. Transfer to Zakai is a normalization change: the object and linearity change, connected by an explicit formula.

Relationships to Other Abstractions¶

Current abstraction Kushner–Stratonovich Equation Domain-specific

Parents (2) — outgoing typed relations

Kushner–Stratonovich Equation is a kind of Bayesian Updating Prime

Bayesian Updating. The node is a strict continuous-time specialization of evidence-conditioned posterior revision, with innovation supplying the incremental likelihood information.
Kushner–Stratonovich Equation includes Stochastic Process Prime

Stochastic Process. Strict composition / part-of: the hidden signal, observation, innovation, and conditional-law trajectory are constitutive indexed random processes, even though the equation is not merely a process species.

Hierarchy paths (4) — routes to 3 parentless roots

Kushner–Stratonovich Equation → is a kind of Bayesian Updating → is a kind of Inductive Reasoning

Show alternative paths (3)

Neighborhood in Abstraction Space¶

Kushner–Stratonovich Equation sits in a moderately populated region (50^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Statistical Learning & Model Failure Modes (41 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-25