Parallel processing for object oriented robotic simulation of tracheal-oesophageal puncture

Abstract

An approach in advancing the field of medical robotics is via a more efficient preparation of robotic surgical intervention using accurate simulation. Our intra-operative virtual system for robotic tracheal-oesophageal puncture is modeled using an object oriented method. The interaction between rigid tools and flexible organs is simulated with finite element method codes integrated within the framework of multi-body dynamics codes. We investigate LogP and MLogP models for multi-core CPU computing, and propose Topology Independent Model (TIN) model for a heterogeneous computing system, which mimics parallel processing with multi-core CPU and GPU, to provide the required computational power for tool-tissue simulation.