中文摘要：

触／力觉交互技术应用环境的复杂化，对复杂触／力觉场景的研究提出了新的挑战．文中以虚拟现实牙科手术训练系统为背景，研究基于速度驱动的复杂场景多层级力觉交互算法．文中采用CATIA对下牙列大量数据进行网格简化，建立了下牙列场景的层次细节（Levels Of Detail，LOD）模型；基于人类触觉感知的精度随交互速度的变化规律，设计了速度驱动的LOD模型触发机制；提出了虚拟工具化身SCP（Surface Contact Point）层级映射算法，保证了力觉交互设备的稳定性，并发现了映射约束线段的长度是保证力觉逼真性及实时性协调的关键变量；最后，设计了针对一般几何体和复杂曲面的实验，量化评价文中算法的逼真性及实时性，证明了将人手运动速度作为场景模型复杂度切换驱动条件的有效性．

英文摘要：

In haptic interaction technology, new challenges have been put forward on complex scene rendering. Based on the requirements of a dental training system, velocity driven levels of detail （LOD） haptic rendering algorithm for complex scenes is proposed. Firstly, a large amount of data is simplified with CATIA to realize level-of-detailed modeling for the teeth data. Second- ly, based on analysis of the relationship between the perceptible features and the interaction velocity, velocity-driven LOD force rendering algorithm is proposed as switching criteria between discrete model levels. Thirdly, a surface contact point （SCP） mapping algorithm is developed to ensure the continuous change of virtual force signal at switch time between two successive LOD models, which ensures the stability of haptic rendering. Finally, three experiments based on planes, spheres and complex teeth model are carried out to validate the fidelity and stability of the proposed algorithm. The experiments results illustrate that the algorithm can stably render scenes with large numbers of triangles under various interaction velocities.