论文部分内容阅读
针对数据手套校正过程中真实人手姿态获取困难、校正模型简单的问题,提出一种基于真实抓取经验的校正数据库构造方法.该算法根据人手生理结构和手套传感器布局建立具有更高运动逼真度的虚拟手运动学模型,实现更为灵活的拇指运动和软手掌效果;在不需要任何外部硬件设备的情况下,将真实抓取经验与最优化理论结合,寻找针对给定三维物体及操作任务的最优抓取姿态,从而构造真实人手姿态数据库;利用最小二乘拟合和闭环迭代算法获得精确的数据手套传感器输出模型.实验结果表明:所提出的姿态构造方法能有效提高数据手套校正效率,且操作过程简单,易于实现自动化校正,适合实际工程应用.
Aimed at the difficulty of obtaining the real human hand posture during the data glove calibration and the problem of the simple calibration model, a calibration database construction method based on real crawling experience is proposed. The algorithm builds a database with higher motion fidelity according to the human physiological structure and the glove sensor layout Virtual hand kinematics model, to achieve more flexible thumb movement and soft palm effect; without any external hardware devices, the real crawling experience and optimization theory combined to find for a given three-dimensional objects and operational tasks The optimal attitude of the gripper to construct a real human gesture database.The exact model of the data glove sensor output is obtained by least square fitting and closed-loop iterative algorithm.The experimental results show that the proposed pose construction method can effectively improve the data glove calibration efficiency, And the operation is simple, easy to automate calibration, suitable for practical engineering applications.