Learning from Demonstration: Dynamical Movement Primitives Based Reusable Suturing Skill Modelling Method

Abstract

Establishing a surgical skill model plays an important role for automatic robotic assisted surgery, besides that it is of great value to evaluate physicians' operational skills. In this paper, superficial tissue suture is selected as the modeling object, and suture skill learning and modeling research is conducted. In order to solve the problem of conventional models with poor migration ability for applications in new scenes, a demonstration-decomposition-modeling suture skill learning and modeling framework is proposed. The demonstration suture process is decomposed into sub-processes that repeats in various suture paradigms. The dynamic movement primitives (DMPs)method is introduced to uniformly model the suture trajectory of each sub-process. The superficial suture operation demonstration and acquisition platform is built, and the modeling ability of the proposed method and the adaptability to new scenes are preliminarily verified.