数据分割与注册:智能化多传感训练系统范例研究

Data segmentation and registration:discussion on model of an adaptive multi-sensor training device

针对多传感器系统的数据融合问题,以自适应性肢体康复训练系统为例,对数据分割与注册进行了研究.对于数据分割中的时间序列对准问题,采用三次样条拟合的方法加以解决;对于数据注册中的形式统一化问题,将实际测量参数经特征提取、适应性建模、预测及自评估,变为特征变量和相应的自评估置信概率表述,其他档案性信息和非模拟量参数则通过模糊化处理变为相同形式的表述,再进一步转换为统一的符号形式,即可解决.仿真结果证明了该融合决策方案的有效性.

For the multi-sensor system＇s data fusion problem,data segmentation and registration by self-adaptive physical rehabilitation training system as an example was discussed.Cubic spline fitting method is used to solve the segmentation of time series data alignment problem.In order to solve the format unification,feature extraction,adaptive modeling,prediction and self-assessment are used in characteristic variables and the corresponding confidence probability of self-assessment statements,other files of information and non-analog processing parameters are changed by the same form of expression and then into unified form.The effectiveness of the method was proved with testing results.