摘要
采用基于Vega和Creator的虚拟现实仿真机制和人机交互设备,构建某型工程车训练仿真系统虚拟操作训练环境。根据位置跟踪传感器跟踪受训人员手的位置和姿态,使用数据手套获取受训人员手部各关节数据,驱动虚拟手实现人机交互,完成虚拟操作训练任务,并采用前馈神经网络模式识别法和空间位姿判别法,分别对静态手势和动态操作命令进行识别。
This training simulation system of certain engineering vehicle is constructed by using the VR simulation mechanism based on Vega and Creator with the devices for VR man-machine interaction. The system drives the virtual hands to accomplish the training tasks in the virtual manipulation environment by tracking the trainee' s hand positions and gestures based on Motion Tracking Device and acquiring the da- ta of the trainee' s hand joints based on Data Glove in real world. The mode recognition of Pre-feedback Neural Network based on DG and the space positions and gestures discriminance based on the MTD are used to recognize the static hand gestures and the dynamic hand commands respectively. The efficiency of the presented methods is validated by applying it to the simulation training for a certain engineering vehicle.
出处
《军械工程学院学报》
2009年第6期67-71,共5页
Journal of Ordnance Engineering College
基金
项目来源:江苏省自然科学基金资助项目(BK2006004)
关键词
工程车训练仿真系统
手势识别
B—P神经网络
虚拟操作.
training simulation system of engineering vehicle
hand gesture recognition
B-P neural network
virtual manipulation