IEEE J Biomed Health Inform. 2015 Jul;19(4):1443-50. doi: 10.1109/JBHI.2015.2392796. Epub 2015 Jan 23.
We present a flexible, easy-to-expand digital signal processing method for detecting heart rate (HR) for cardiac vibration signals of fiber Bragg grating (FBG) sensor. The FBG-based method of measuring HR is possible to use during the magnetic resonance imaging procedure, which is its unique advantage. Our goal was to design a detection method with plurality of parameters and to subject these parameters to genetic algorithm optimization technique. In effect, we arrived at a method that is well able to deal with much distorted signals with low SNR. We proved that the method we developed allows automatic adjustment to the shape of the waves of signal carrying useful information about the moments of heartbeat. Thus, we can easily adapt our technique to the analysis of signals, which contains information on HR, from sensors employing different techniques of strain detection. The proposed method has the capabilities of analyzing signals in semi-real-time (online) with beat-to-beat resolution, significantly low delay, and negligible computational power requirements. We verified our method on recordings in a group of seven subjects. Verification included over 6000 heartbeats (82 min 47 s of recordings). The root-mean-square error of our method does not exceed 6.0 bpm.
我们提出了一种灵活、易于扩展的数字信号处理方法，用于检测光纤布拉格光栅（FBG）传感器的心脏振动信号的心率（HR）。基于 FBG 的 HR 测量方法有可能在磁共振成像过程中使用，这是其独特的优势。我们的目标是设计一种具有多种参数的检测方法，并将这些参数应用于遗传算法优化技术。实际上，我们得到了一种能够很好地处理具有低 SNR 的失真信号的方法。我们证明，我们开发的方法能够自动调整携带有关心跳时刻的有用信息的信号波形。因此，我们可以轻松地将我们的技术应用于分析包含 HR 信息的来自使用不同应变检测技术的传感器的信号。所提出的方法具有以半实时（在线）方式进行分析的能力，具有逐拍分辨率、显著低延迟和可忽略的计算功率要求。我们在七名受试者的一组记录中验证了我们的方法。验证包括超过 6000 次心跳（82 分钟 47 秒的记录）。我们的方法的均方根误差不超过 6.0 bpm。
