Active-Model-Driven Online Prediction of Large Curvature Bending for Flexible Needle in Percutaneous Interventions Within Soft Tissues

In percutaneous intervention using flexible needle, larger bending means larger reachable space, but accurately predicting the bending behavior of flexible needle is challenging due to the nonlinear large deformation. This has become a critical factor affecting the precision of operation. In this study, an active-model-driven approach is proposed to predict the deformation of the flexible needle when confronted with large curvature bending during insertions. Initially, we analyze the needle-tissue interaction and establish mechanics-based models to simulate the interaction forces throughout the insertion process. Subsequently, a nonlinear mechanical model, denoted as the reference model, is proposed to describe the large bending of the needle. It is further augmented with an active-model-driven approach to actively estimate the modeling errors and make compensations. Percutaneous intervention experiments are conducted on multi-layer tissues and ex vivo porcine brains. The results show that the proposed method can accurately predict the large curvature bending of the flexible needle.