Sequential evaluation of a robotic dry-bone resection system using virtual model control with integral control: an exploratory benchtop study

Abstract Virtual model control (VMC) provides compliant trajectory execution, but sustained contact loads can produce residual positional error. This study evaluated the sequential development of VMC augmented with integral control for powered dry-bone resection and assessed the effects of end-effector stabilisation, spatial calibration and operating parameters. A seven-degree-of-freedom Franka Research 3 robot was commanded using VMC with integral control and carried a 4-mm diamond-tipped arthroscopic burr through a programmed 10 × 10 mm serpentine trajectory in 20 pounds-per-cubic-foot synthetic bone. Four sequential configurations were assessed: an original handle-based mount, an anthropomorphically designed shaft-stabilising mount, robot-frame calibration and workpiece registration. Each configuration comprised three resections. Performance was graded by one non-blinded assessor using a study-specific twelve-point score across burr-tip oscillation and X-, Y- and Z-axis deviation, with lower scores indicating better performance. Radial depth of cut, engagement depth, trajectory velocity and burr angle were then screened to identify candidate operating parameters. Mean total score decreased sequentially from 11.7/12 with the original mount to 7.7/12 after shaft stabilisation, 5.7/12 after robot-frame calibration and 4.0/12 after workpiece registration. The lowest observed score was maintained at an RDOC between 0.10 and 0.25 mm, a measured depth between 1.52 and 1.89 mm, velocity between 1.5 and 2.5 mm/s and a burr angle between 30 and 45 degrees. Mechanical stabilisation and spatial calibration were associated with progressively better resection performance using VMC with integral control. This exploratory benchtop study identified candidate operating ranges for further assessment in higher-fidelity models.