Bonnet Vincent, Venture Gentiane
IEEE Trans Neural Syst Rehabil Eng. 2015 Jul;23(4):628-35. doi: 10.1109/TNSRE.2015.2405087. Epub 2015 Mar 2.
This study aimed at developing and evaluating a new method for the fast and reliable identification of body segment inertial parameters with a planar model using affordable sensors. A Kinect sensor, with a new marker-based tracking system, and a Wii balance board were used as an affordable and portable motion capture system. A set of optimal exciting motions was used in a biofeedback interface to identify the body segment parameters. The method was validated with 12 subjects performing various standardized motions. The same dynamometric quantities estimated both with the proposed system and, as a reference, with a laboratory grade force-plate were compared. The results showed that the proposed method could successfully estimate the resultant moment and the vertical ground reaction force (rms errors less than 8 Nm and 12 N, respectively). Finally, when local segment values were artificially varied, the proposed method was able to detect and estimate the additional masses accurately and with an error of less than 0.5 Kg, contrary to values generated with commonly used anthropometric tables.
本研究旨在开发并评估一种新方法，该方法使用价格实惠的传感器，通过平面模型快速可靠地识别身体各节段的惯性参数。配备新型基于标记的跟踪系统的Kinect传感器和Wii平衡板被用作价格实惠且便携的运动捕捉系统。在生物反馈界面中使用一组最佳激励动作来识别身体各节段参数。该方法通过12名受试者进行各种标准化动作进行了验证。比较了使用所提出的系统以及作为参考的实验室级测力板估计的相同测力参数。结果表明，所提出的方法能够成功估计合力矩和垂直地面反作用力（均方根误差分别小于8 Nm和12 N）。最后，当人为改变局部节段值时，与常用人体测量表生成的值相反，所提出的方法能够准确检测并估计附加质量，误差小于0.5 Kg。
