Multilevel Fuzzy Control Based on Force Information in Robot-Assisted Decompressive Laminectomy.

The lumbar spinal stenosis (LSS) is a kind of orthopedic disease which causes a series of neurological symptom. Vertebral lamina grinding operation is a key procedure in decompressive laminectomy for LSS treatment. With the help of image-guided navigation system, the robot-assisted technology is applied to reduce the burdens on surgeon and improve the accuracy of the operation. This paper proposes a multilevel fuzzy control based on force information in the robot-assisted decompressive laminectomy to improve the quality and the robotic dynamic performance in surgical operation. The controlled grinding path is planned in the medical images after 3D reconstruction, and the mapping between robot and images is realized by navigation registration. Multilevel fuzzy controller is used to adjust the feed rate to keep the grinding force stable. As the vertebral lamina contains different components according to the anatomy, it has different mechanical properties as the main reason causing the fluctuation of force. A feature extraction method for texture recognition of bone is introduced to improve the accuracy of component classification. When the inner cortical bone is reached, the feeding operation needs to stop to avoid penetration into spinal cord and damage to the spinal nerves. Experiments are conducted to evaluate the dynamic stabilities of the control system and state recognition.