Chang Nai-Fu, Chiang Cheng-Yi, Chen Tung-Chien, Chen Liang-Gee
School of Veterinary Medicine, National Taiwan University, Taipei, Taiwan.
Annu Int Conf IEEE Eng Med Biol Soc. 2011;2011:7091-4. doi: 10.1109/IEMBS.2011.6091792.
On-chip implementation of Hilbert-Huang transform (HHT) has great impact to analyze the non-linear and non-stationary biomedical signals on wearable or implantable sensors for the real-time applications. Cubic spline interpolation (CSI) consumes the most computation in HHT, and is the key component for the HHT processor. In tradition, CSI in HHT is usually performed after the collection of a large window of signals, and the long latency violates the realtime requirement of the applications. In this work, we propose to keep processing the incoming signals on-line with small and overlapped data windows without sacrificing the interpolation accuracy. 58% multiplication and 73% division of CSI are saved after the data reuse between the data windows.
希尔伯特-黄变换（HHT）的片上实现对于分析可穿戴或植入式传感器上的非线性和非平稳生物医学信号以用于实时应用具有重大影响。三次样条插值（CSI）在HHT中消耗的计算量最大，并且是HHT处理器的关键组件。传统上，HHT中的CSI通常在收集一大窗口信号之后执行，而长延迟违反了应用的实时要求。在这项工作中，我们建议在不牺牲插值精度的情况下，使用小的重叠数据窗口对传入信号进行在线持续处理。在数据窗口之间进行数据复用后，CSI节省了58%的乘法运算和73%的除法运算。
