Smooth path planning for a biologically-inspired neurosurgical probe

Percutaneous intervention involves the insertion of needles to specific locations inside the human body, to perform a variety of surgical procedures. Percutaneous procedures are becoming the preferred choice for many neurosurgeons, due to the additional benefits they provide over conventional open neurosurgery. A neurosurgical flexible and steerable probe named STING is currently being developed for accessing deep brain lesions following curvilinear paths. In this paper, we present a path planning method for generating pre-operative paths for this neurosurgical flexible probe. Since the flexible probe is modeled as a nonholonomic system, a deterministic continuous curvature path planning scheme capable of avoiding obstacles is developed for smooth steering of its tip. Multiple paths are generated by varying arrival angle at the targeted lesion and a path optimization approach is then formulated, with the aim to minimize damage to the tissue (i.e. shortest path) and the risk to the patient (obstacle avoidance). Simulation results are reported using the risk-map generated for a coronal slice of the brain, which confirms the successful design of a path planning scheme that satisfies the nonholonomic constraints of the neurosurgical probe.