A state-space model tracks hidden internal variables to describe how a system evolves

A state-space model tracks hidden internal variables to describe how a system evolves
Image: Intellect (generated)

In short: Engineers and economists often describe a system by following a set of internal state variables instead of only watching its inputs and outputs. This state-space approach turns system behaviour into matrix algebra, copes with many inputs and outputs at once, and supports tools such as the Kalman filter, which estimates quantities that cannot be measured directly.

A state-space representation models a physical system with state variables. These variables capture the condition of the system at any moment and change according to their present values and the inputs applied. Outputs depend on the states and, in some cases, directly on the inputs. The axes of a geometric space can be the state variables themselves, and the condition of the system at a given time is then a single point, called the state vector.

For linear systems whose behaviour does not change over time and that have a finite number of parts, the equations can be arranged neatly into matrices. Compared with the frequency-domain method based on Laplace transforms, this layout handles several inputs and outputs naturally and does not require starting from zero initial conditions. The minimum number of state variables often matches the order of the governing differential equation, though not always. In electrical circuits it frequently equals the count of energy-storing parts such as capacitors and inductors.

The framework also supports practical questions about control. Whether a system is stable can be read from the eigenvalues of its matrix, and whether it can be steered or inferred from its outputs can be tested with rank conditions. Converting a model to a transfer function can discard some internal information, so a description that looks stable from outside may hide an unstable internal state. Beyond engineering, these models appear in econometrics, statistics, computer science and neuroscience. Economists use them to split a time series into trend and cycle, and to estimate output measures that are only partly observed. Many such uses depend on the Kalman filter or a state observer, which infer current hidden states from earlier observations.

Opens the Intellect+ feed with this card first. Swipe up for the next one.

Sources

To cite this card, link to https://intellectplus.co.uk/app/science-state-space-representation.

How we source, check and correct cards: Editorial standards.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

Finding something worth knowing…