Abstract

Virtual surgery simulators are an emerging training method for medical specialists. A number of techniques have previously been proposed for simulating tissue for these simulators, one of which is the mass-spring model.

This paper aims to improve upon the perceived realism of virtual simulator systems by redesigning the underlying physics algorithms so that accurate simulations with full and smooth haptic interaction can be performed.

We present a novel mass-spring model algorithm designed to accurately simulate soft human organ tissue, both through visual (graphic) and haptic perception (touch). The algorithm is implemented as a massively parallel algorithm running on theNVIDIA CUDA computing architecture, allowing it to simulate human tissue at a rate of at least 1000 Hz, the minimal rate at which we can ensure smooth haptic interaction. The mass-spring model is integrated as part of VICTAR, or Virtual Competence Training Area, an extendable soft- ware framework designed for virtual surgical training. VICTAR features haptic device handling and a scripting engine that allows for easy prototyping of certain medi- cal scenarios and also serves as the primary method for benchmarking. The results shows that the presented implementation is capable of sustaining rates above the requirements and is therefore fast enough to ensure smooth haptic interaction.

Additional Resources

1. The paper