Polar Decomposition Calculator
Compute the polar decomposition of 2x2 and 3x3 real matrices into unitary/orthogonal (U) and positive semi-definite (P) factors with step-by-step math.
What is Polar Decomposition of a Matrix?
In linear algebra, the polar decomposition of a square matrix $A$ is a fundamental matrix factorization analogous to the polar form of a complex number $z = r e^{i\theta}$.
Just as any complex number can be expressed as the product of a non-negative magnitude $r \ge 0$ and a unit rotation $e^{i\theta}$, any real square matrix $A \in \mathbb{R}^{n \times n}$ can be factored into:
$$A = U \cdot P \quad \text{(Right Polar Decomposition)}$$
or
$$A = P' \cdot U \quad \text{(Left Polar Decomposition)}$$
where:
- $U$ is an orthogonal matrix (or unitary matrix in the complex case), meaning $U^T U = I$ and $\det(U) = \pm 1$. Geometrically, $U$ represents a pure rotation or reflection.
- $P = \sqrt{A^T A}$ is a symmetric positive semi-definite matrix representing pure stretching / scaling along orthogonal principal axes.
- $P' = \sqrt{A A^T} = U P U^T$ is the corresponding left positive semi-definite stretch factor.
How to Calculate Polar Decomposition
The most numerically stable and standard method to calculate the polar decomposition is via Singular Value Decomposition (SVD):
- Singular Value Decomposition (SVD): Compute the SVD of $A = W \Sigma V^T$, where $W$ and $V$ are orthogonal matrices and $\Sigma = \operatorname{diag}(\sigma_1, \dots, \sigma_n)$ contains the non-negative singular values $\sigma_i \ge 0$.
- Unitary Factor ($U$): Set $U = W V^T$.
- Right Stretch Factor ($P$): Set $P = V \Sigma V^T = \sqrt{A^T A}$.
- Left Stretch Factor ($P'$): Set $P' = W \Sigma W^T = \sqrt{A A^T}$.
Geometric Interpretation in Mechanics and Graphics
Polar decomposition plays a critical role in continuum mechanics, robotics, computer graphics, and elasticity theory. When a physical body undergoes deformation described by deformation gradient tensor $F$:
- $F = R U$: The deformation is decomposed into pure stretch $U$ followed by rigid-body rotation $R$.
- $F = V R$: The deformation is decomposed into rigid-body rotation $R$ followed by spatial stretch $V$.
Related Matrix and Linear Algebra Tools
Explore other matrix factorization and linear algebra calculators:
- Diagonalize Matrix Calculator: Factorize square matrices into $P D P^{-1}$.
- Cholesky Decomposition Calculator: Factor symmetric positive-definite matrices into $L L^T$.
- Characteristic Polynomial Calculator: Find matrix eigenvalues and trace.
Frequently Asked Questions
Is the polar decomposition unique?
If $A$ is invertible (non-singular, $\det(A) \ne 0$), the polar decomposition $A = UP$ is strictly unique. If $A$ is singular, the positive semi-definite matrix $P = \sqrt{A^T A}$ is still unique, but the orthogonal factor $U$ is not uniquely determined on the null space of $A$.
How does polar decomposition relate to SVD?
Polar decomposition is directly derived from SVD. Given $A = W \Sigma V^T$, the polar factors are $U = W V^T$ and $P = V \Sigma V^T$. SVD separates both input and output rotation bases, whereas polar decomposition isolates a single overall rotation and a symmetric stretch.
What is the difference between Right and Left Polar Decomposition?
Right polar decomposition writes $A = UP$, where stretching ($P$) occurs before rotation ($U$). Left polar decomposition writes $A = P'U$, where rotation ($U$) occurs before stretching ($P'$). The rotation matrix $U$ is identical in both decompositions, and $P' = U P U^T$.
What are the eigenvalues of matrix P?
The eigenvalues of $P = \sqrt{A^T A}$ are exactly the singular values $\sigma_1, \sigma_2, \dots, \sigma_n$ of the original matrix $A$. All eigenvalues of $P$ are real and non-negative.