Spatial Curve Approximation for Pneumatic Actuators: A Comparative Study of Spiral and Spline Models

Soft robotics relies heavily on precise movement control, particularly through pneumatic actuators, to achieve complex, bio-inspired motions. This paper investigates multiple curve approximation methods to model the kinematic behavior of pneumatic networks (pneunets) that produce bending motions. Using experimental data from pneunets with varied wall thicknesses, we compare the accuracy of logarithmic, Archimedean, and quadratic spirals, as well as cubic B-splines and Hermite-splines, in approximating the pneunets’ spatial curves. Each approach exhibits distinct curvature characteristics, influencing how accurately it represents the pneunet’s bending trajectory. Results indicate that the quadratic spiral provides the most accurate fit with minimal deviation, while B-splines and Hermite-splines also achieve reliable approximations when control points are optimized. The quadratic spiral’s adaptability across different designs makes it a valuable straightforward model for soft robotic applications, effectively capturing complex bending behavior with precision and efficiency.