Driving flexible needles in 3D tissue without obstacles

The bevel-tip flexible needle is an improvement of the clinical puncture needle and has the potential to reduce surgery traumas and improve puncture accuracy. Due to the nonholonomic kinematics model, it is difficult to drive flexible needles to a target precisely. In this paper, a driving method is proposed to steer flexible needles to a target in 3-D environment without obstacles. In addition, a real-time closed-loop control system is set up based on the method. The proposed method needs a priori knowledge of model parameter, which is obtained by offline identifying and the needle tip's position and attitude. We obtain the tip's position with an image-guided system and employ UKF (Unscented Kalman Filter) to estimate the needle tip's attitude. Simulations verify that the driving method is precise and robust. The tip error will be confined to 0.5mm when the insertion distance is 120mm.