Exploration of an admittance control method for multi-segment spinal motion loading.

BACKGROUND: In vitro testing is a fundamental approach for advancing spinal biomechanics research. However, existing loading methods still exhibit notable limitations in physiological realism and motion controllability. This study introduces a spinal loading method that integrates robotic admittance control with dynamic movement primitives (DMPs) to enable motion loading of multi-segment spines, and provides a preliminary assessment of its reliability. METHODS: A robotic testing system was developed, comprising a six-degree-of-freedom robotic arm integrated with a six-axis force sensor. Spinal traction trajectories were acquired with reference to anatomical planes, modeled, and subsequently reproduced using DMPs. The reconstructed trajectories were compared with the reference data to assess reproduction accuracy. The proposed method was further validated on ovine thoracolumbar specimens (T12-L3) under flexion-extension and lateral bending motions, and the measured ranges of motion (ROM) were compared with values reported in previous studies. FINDINGS: A spinal loading method based on robotic admittance control combined with DMPs was successfully established. The experimental results demonstrated that the mean error between the demonstrated and reproduced trajectories was less than 2.5 mm, and the measured range of motion showed no significant difference compared with previously reported data (P < 0.05). INTERPRETATION: The proposed method accurately reproduces physiological spinal motion characteristics, demonstrating its feasibility and validity for in vitro studies of three-dimensional spinal biomechanics.