The expression x′y″ − y′x″ combines first and second derivatives to capture how the direction of a parametrized path changes. The denominator, (x′² + y′²)^(3/2), accounts for the rate associated with the chosen parameter. Taking the magnitude produces a nonnegative measure, so sharper directional changes correspond to larger curvature values.
Arc length measures distance traveled along the curve rather than change in an arbitrary parameter. Describing the unit tangent vector with respect to arc length therefore focuses on geometric turning along the path itself. This makes curvature useful for comparing different paths and distinguishing changes in shape from differences caused only by how a curve is parametrized.
Zero curvature corresponds to a straight trajectory because its direction does not bend. Constant curvature describes circular motion, giving the path a uniform turning behavior. When curvature varies from point to point, the path bends by different amounts along its length, allowing mathematical analysis to identify tighter and gentler sections of the same trajectory.
For coordinates written as x and y functions of a parameter, first compute the first derivatives x′ and y′, then the second derivatives x″ and y″. Substitute them into κ = |x′y″ − y′x″|/(x′² + y′²)^(3/2). Evaluating this expression along the path reveals where the trajectory bends more sharply or more gently.
Curvature provides a local measure of turning that can be evaluated at corresponding points or sections of different paths. A larger value identifies a tighter bend, while a smaller value indicates gentler turning. This supports geometric comparison beyond overall appearance, because two shapes can be examined according to how their bending changes along their trajectories.
In robotics and computer graphics, curvature helps evaluate turning behavior when paths must be designed or analyzed. It can identify abrupt or tight bends and support smooth path design. The same geometric measure connects trajectory analysis with practical planning and visual modeling, where the quality of a path depends on how its direction changes.