Dataset of simulated deformations of a three-fiber soft robotic arm modeled based on morphoelasticity and active filament theory

Dataset Description This dataset contains simulation data of a three-fiber soft robotic arm generated using a custom Julia package (Paper: https://www.sciencedirect.com/science/article/pii/S0022509622001181 , GitHub: https://github.com/LivingMatterLab/ActiveFilaments.jl ). Each simulation defines a set of fiber activation inputs γi and solves a boundary value problem to obtain the robot’s centerline configuration (x, y, z) along the spatial coordinate z ∈ [0, L]. A total of N = 1,000,000 simulations were performed using random activation triplets within the physical range γi ∈ [-1.67, 0]. The resulting dataset provides a large-scale mapping from actuation inputs to corresponding robot shapes. Contents The file load_data.py loads the following data: activations_gamma.npz: Array of shape N × 3, containing the activation inputs γ1, γ2, γ3. centerline_r.npz: Array of shape N × 100 × 3, giving the computed centerline positions in Cartesian coordinates (x, y, z) at 100 discretized spatial points. coordinate_z.npz: Array of shape N × 100, specifying the discretized spatial coordinate z for each simulation. Usage The data has been employed in the following publications: Shape-Space Graphs: Fast and Collision-Free Path Planning for Soft RobotsThe dataset serves as the shape-library for constructing shape-space graphs.Paper: https://arxiv.org/abs/2510.03547 GitHub: https://github.com/cayobro/shape-space-graphs.git Infinite-Dimensional Closed-Loop Inverse Kinematics for Soft Robots via Neural OperatorsThe dataset is used for training and validation of the neural operator model.Paper: https://arxiv.org/abs/2602.18655 GitHub: https://github.com/cayobro/soft-robot-CLIK-via-neural-operators.git Citation If you use this dataset, please consider citing the above works.