基于实测数据的软组织形变系统辨识方法

Identification Method of Soft Tissue Deformation System Based on Measured Data

为了实现虚拟手术场景的视觉真实感和触觉真实感,需要建立准确的软组织形变模型。从软组织真实形变角度出发结合系统辨识理论来对软组织进行定量建模。考虑到软组织的各向异性,分别采集软组织的水平形变数据和垂直形变数据,并利用LS-SVR建立软组织应力与应变之间的关系,从而完成对软组织水平方向和垂直方向的形变建模。同时通过对软组织表面的点进行固定采样,并结合所建立的模型可以实现虚拟手术系统输入力输出形变位移的导纳控制。将系统辨识输出的形变结果与真实的形变数据进行对比,辨识结果具有较高的准确性,可以满足虚拟手术系统真实性的要求。

In order to realize the realistic visual sense and touch sense in virtual surgery,we need established accurate soft tissue deformation model.From the view of real soft tissue deformation,this article combines the theory of system identification to quantitatively model soft tissue.Considering the anisotropy of soft tissue,we collected horizontal and vertical deformation data of soft tissue respectively,and established the relationship between stress and strain of soft tissue by using LS-SVR,so as to complete the deformation modeling of soft tissue in horizontal and vertical directions.At the same time,we can realize the admittance control of the input and output deformation in virtual surgery system by sampling the points on the soft tissue surface.By comparing the deformation results of the system identification output with the real deformation data,the identification results have higher accuracy,and can satisfy the requirements of the virtual surgery system.