Spinal stenosis is a common orthopedic disease that can cause a range of neurological symptoms. Laminectomy is a common treatment for spinal stenosis; however, due to the proximity of the surgical area to the vulnerable spinal nerves, the procedure requires high precision and is time-consuming, thus only experienced surgeons can perform it. With technological advancements, robot-assisted surgery is becoming increasingly accepted. To prevent the robotic end effector from penetrating the lamina and damaging nerves, this study proposes a real-time grinding depth monitoring method based on a grinding force model to ensure surgical safety. Specifically, firstly, a cutting force model for the spherical grinding head at different depths is established based on its geometry; then, bovine bone is used as the grinding experimental material, and force signals from the robotic end effector are collected during the experiment. Unknown parameters in the cutting force model are identified using a particle swarm optimization algorithm; finally, a model of the relationship between robotic grinding depth and grinding force is established through multiple experiments. Experimental results show that when the robot grinding depth is [0.8, 1.2] mm, the grinding depth prediction accuracy based on grinding force can reach ±0.1 mm; when the depth increases to [1.2, 1.6] mm, the prediction accuracy is ±0.2 mm.
The real-time monitoring method for robotic grinding depth based on a grinding force model proposed by this research institute has significant clinical application value. It enables precise intraoperative monitoring of the patient's condition during robot-assisted laminectomy, ensuring the safety of the robotic surgery. Furthermore, this technology can be effectively applied to robot-assisted neurosurgery, joint replacement, and other surgeries.
