非接触式生命信号特征提取方法研究

发布时间：2018-05-01 11:11

本文选题：非接触式 + 时分短时窗　；参考：《南京理工大学》2017年硕士论文

【摘要】：非接触式生命信号检测技术是利用雷达远距离测量人体呼吸、心跳特征的新兴技术。近年来,随着检测基础理论的完善,硬件成本的降低,该技术逐渐进入实际应用阶段。本文正是在这样的背景下,针对实际应用中的家庭监护、医疗检测两种热点场景,分别研究了两类生命信号特征提取方法:(1)面向家庭监护应用的实时提取方法。为提高实时性,本方法利用低频载波下的小角度近似解调,直接获取近似生命信号。通过时域寻峰法对呼吸频率进行实时提取。针对传统方法提取心率延时较长的缺点,提出基于时分短时窗的心率快速提取算法。并通过仿真和实测验证,该算法单次测量延时为2-5s,相比传统算法实时性显著提高。(2)面向医疗检测应用的高精度提取方法。为提高精度,本方法在高频载波下获取高相位分辨率的雷达基带信号。对实际硬件电路存在的正交失衡、直流偏置的影响,采用椭圆校准算法、拟合圆算法对精密运动信息进行恢复。提出扩展DACM正交解调算法获取生命信号线性波形。设计单摆实验测量了系统的正交失衡因子,同时验证了基带信号预处理算法的必要性、可靠性。采用完全集合经验模态算法(CEEMD)对呼吸、心跳信号进行分离,经与参考值对比,呼吸、心跳速率测量准确率达到98%以上,提取出的心跳时域波形达到心搏间隔可测,为医学诊断提供了具有参考意义的特征信息。对上述两种应用场景,分别搭建了实时提取系统、高精度提取系统,完成了系统的调试工作,并分析两种系统在人体抖动和呼吸谐波干扰下的性能与误差,并得出相应的结论,为后续工作提供了一定思路。
[Abstract]:Non-contact life signal detection technology is a new technology which uses radar to measure human breathing and heartbeat characteristics. In recent years, with the improvement of detection theory and the reduction of hardware cost, the technology has gradually entered the stage of practical application. Under this background, this paper studies two kinds of life signal feature extraction methods: one is the real time extraction method for the family monitoring application, the other is the real time extraction method for the family monitoring application, aiming at the two hot spots of family monitoring and medical detection. In order to improve the real-time performance, the approximate life signal is obtained by using the small angle approximate demodulation under low frequency carrier. The time domain peak finding method was used to extract the respiratory frequency in real time. Aiming at the disadvantage of the traditional method, a fast heart rate extraction algorithm based on time division short time window is proposed. The simulation and experimental results show that the time delay of single measurement is 2-5s. Compared with the traditional algorithm, the real-time performance of the proposed algorithm is significantly improved. (2) A high-precision extraction method for medical detection applications is proposed. In order to improve the accuracy, a high phase resolution radar baseband signal is obtained under high frequency carrier. For the effect of orthogonal imbalance and DC bias in the actual hardware circuit, the elliptical calibration algorithm and the fitting circle algorithm are used to restore the precise motion information. An extended DACM quadrature demodulation algorithm is proposed to obtain linear waveform of life signal. A single pendulum experiment was designed to measure the orthogonal imbalance factor of the system, and the necessity and reliability of the baseband signal preprocessing algorithm were verified. The complete set empirical mode algorithm (CEEMD) was used to separate the respiration and heartbeat signals. Compared with the reference values, the measurement accuracy of respiration and heartbeat rate was over 98%, and the extracted time-domain waveform of heartbeat could be measured. It provides the characteristic information for medical diagnosis. For the two application scenarios mentioned above, the real-time extraction system and the high-precision extraction system are built, the debugging work of the system is completed, and the performance and error of the two systems under the human body jitter and respiratory harmonic interference are analyzed, and the corresponding conclusions are drawn. It provides some ideas for the follow-up work.
【学位授予单位】：南京理工大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.7

【相似文献】