虚拟手术训练系统中人机交互问题的研究

The lack of adequate hands-skill training has long been a great dilemma in medicine. Because of its many advantages and potentials, the visual reality and haptics based surgical simulator is widely regarded as an ideal and feasible solution to this problem. Due to its high level of complexity, surgical simulators, however, are still at their infancy stage and there still exist numerous fundamental challenges and technical difficulties. In this research, we will focus on the human-machine interfacing aspect of surgical simulation and address some major problems. First, we will investigate to use the videos of real surgery as the background, integrate them with simulated surgical scenes, and render them in a seamless manner -- realizing a high-degree of visual fidelity and immersion. Second, new methods will be studied to compensate for the gravity of haptic interface devices. Through structure optimization and controller design, a new mechanism of haptic interfaces with the capability of automatic gravity compensation will be developed. In order to enhance system transparency under the condition of stability, we propose an optimal control approach, which is based on the multi-channel ISS and IOS small-gain theorem. At the end, the expected research results will be used for the development of neuron surgery simulators and their feasibility and effectiveness will be verified.

外科医生实操技能培训的不足一直是困扰医学界一个重大难题。基于虚拟现实和力触觉交互技术的虚拟手术训练系统由于其诸多优点，被业界广泛视为可能破解这一困境的一种解决方案。但由于其高度复杂性，虚拟手术训练系统目前还处于比较简单和初级阶段，还有很多科学问题和技术难题亟待解决。本项目拟重点研究虚拟手术训练系统中人机交互方面的一些问题：首先，以真实手术视频为背景，研究虚实融合增强方法，实现手术视频与模拟手术场景之间的高度融合和一致性呈现，提高视觉反馈的真实感和沉浸感；其次，研究力触觉机械手重力自主调节和平衡机构，并通过结构优化，结合控制算法设计，实现力触觉机械手自主重力补偿的目的；然后，基于多通道ISS和IOS“小增益理论”的思想，研究模拟手术器械与软组织力交互操作的稳定性控制算法和系统透明性优化方法；最后，将拟取得的理论结果应用于研发虚拟脑肿瘤切除手术系统中，以验证其正确性和有效性。

本项目针对虚拟手术中人机交互目前存在的一些主要问题，重点研究了软组织形变模型、软组织三维重建、软组织切割与流血仿真以及视觉呈现等视觉交互方面的主要问题，提出了相关的解决方案或改进算法；在触觉交互方面，研究了适合虚拟手术的专用力触觉机械手设计方法，提出了新的触觉力交互的控制算法；最后以脑膜瘤切除手术为例，研究发了虚拟手术模拟系统实现中的一些问题，研究结果对外科医生的训练和培养具有较重要的意义和应用前景。