Park Wooram, Reed Kyle B, Okamura Allison M, Chirikjian Gregory S
Department of Mechanical Engineering, Johns Hopkins University, 3400 N. Charles St., Baltimore, MD, USA.
IEEE Int Conf Robot Autom. 2010:3703-3708. doi: 10.1109/ROBOT.2010.5509380.
Flexible needles with bevel tips are being developed as useful tools for minimally invasive surgery and percutaneous therapy. When such a needle is inserted into soft tissue, it bends due to the asymmetric geometry of the bevel tip. This insertion with bending is not completely repeatable. We characterize the deviations in needle tip pose (position and orientation) by performing repeated needle insertions into artificial tissue. The base of the needle is pushed at a constant speed without rotating, and the covariance of the distribution of the needle tip pose is computed from experimental data. We develop the closed-form equations to describe how the covariance varies with different model parameters. We estimate the model parameters by matching the closed-form covariance and the experimentally obtained covariance. In this work, we use a needle model modified from a previously developed model with two noise parameters. The modified needle model uses three noise parameters to better capture the stochastic behavior of the needle insertion. The modified needle model provides an improvement of the covariance error from 26.1% to 6.55%.
带有斜角尖端的柔性针正被开发成为微创手术和经皮治疗的有用工具。当将这种针插入软组织时，由于斜角尖端的不对称几何形状，针会发生弯曲。这种带有弯曲的插入操作并非完全可重复。我们通过将针反复插入人造组织来表征针尖姿态（位置和方向）的偏差。针的基部以恒定速度推进而不旋转，并且根据实验数据计算针尖姿态分布的协方差。我们推导了封闭形式的方程来描述协方差如何随不同的模型参数变化。我们通过匹配封闭形式的协方差和实验获得的协方差来估计模型参数。在这项工作中，我们使用了一个从先前开发的模型修改而来的针模型，该模型有两个噪声参数。修改后的针模型使用三个噪声参数来更好地捕捉针插入的随机行为。修改后的针模型将协方差误差从26.1%降低到了6.55%。
