Analytical Safety in Autonomous Surgical Drilling

Analytical safety for fault detection is a well established and well used technique in industrial automation; it has considerable potential to enhance safety in surgical robotic systems. This technique uses multiple inputs and knowledge of plant processes to detect and diagnose a larger range of faults than the conservative safe-design techniques typically used in surgical robot design. In this paper, a sliding mode observer for robust state-estimation and fault reconstruction is designed using system and drilling process models. To account for parametric uncertainty, the fault signals under normal fault-free operating conditions are investigated. A fault-slate is identified when the fault signals lay outside a certain region on the fault signal plane. A second region is identified that signifies pre-contact and post-breakthrough situations. Thus, the scheme is also able to both detect faults and identify stages in the drilling process.