Human-computer interaction on virtual reality-based training system for vascular interventional surgery

Abstract

Background and objective

Currently, precision control and interaction between virtual hand models and ultrafine, ultra-long flexible guidewires in virtual vascular interventional surgery training systems still remain challenging.

Methods

To tackle this issue, this study utilized a hybrid approach combining Cosserat rod theory with quaternions to develop a model for ultra-long flexible guidewires. Through the implementation of a spatial hash-based continuous collision detection (CCD) algorithm, the system achieved precise collision detection between the guidewire and blood vessels. Additionally, adhesive collision particles were integrated into the fingers of the virtual hand model involved in interactions, facilitating the simulation of intervention tasks such as grasping and delivering. CCD technology, in conjunction with extended bounding volume, was employed in the blood vessel model to prevent tunneling effects resulting from rapid hand manipulations.

Results

Experiments were conducted to assess the picking, delivery, and consistency of delivery distance, showcasing the alignment of manipulation between the virtual hand models and real hands when handling the guidewire. The virtual hand model successfully navigated the flexible guidewire model into vessels curved at angles of 30°, 60°, 90°, and 120°, achieving an average response time of 12.64 ms. Moreover, across vessel models curved at various angles, the average disparity between the delivery distance along the x-axis by the hand in a real environment and the guidewire's delivery distance within the virtual vessel model was approximately 3.71 mm, showcasing a high level of smoothness and stability in the interaction between the hand model and the guidewire model.

Conclusions

Finally, within the virtual system, the successful navigation of the hand delivering the guidewire through the femoral artery and radial artery towards the heart further demonstrates the excellent interaction performance between the virtual hand model and the ultrafine, ultra-long flexible guidewires. This success provides both theoretical and experimental support for the interactive training of virtual hand models and guidewires within virtual surgical training systems.