Multi-sensor dataset of a tendon-driven continuum robot in dynamic motion

<p><br></p><p dir="ltr">We present a multi-sensor dataset of a single-segment tendon-driven continuum robot undergoing quasi-static motions, dynamic trajectories, and contact interactions with a tracked object in the workspace.</p><p dir="ltr">Six sensing modalities are recorded synchronously: optical motion capture of five disk poses along the backbone, fiber-optic distributed strain sensing (with shape reconstruction), tendon tension and displacement, base force--torque, and an external object pose and contact force.</p><p dir="ltr">The dataset is organized into three subsets targeting distinct research tasks: quasi-static configurations for mechanical parameter identification and static model validation; dynamic trajectories (planar sweeps, star patterns, circular paths, and Lissajous curves at two actuation speeds) for dynamics modeling, state estimation, and learning-based methods; and dynamic motions with simultaneous contact measurements for contact interaction analysis.</p><p dir="ltr">Both raw sensor recordings and processed data (low-pass filtered at 15 Hz, uniformly resampled at 100 Hz, and time-aligned) are provided in a consistent file structure with documented variable naming.</p><p dir="ltr">Dataset quality is characterized through cross-modal validation, including pairwise cross-correlation of all sensor streams and force--kinematics consistency checks.</p><p dir="ltr">The robot platform is assembled from commercially available components with a fully documented design.</p>