生物雷达非接触检测中呼吸和心跳信号分离算法研究

发布时间：2018-05-31 07:53

本文选题：生物雷达 + LMS算法　；参考：《第四军医大学》2012年硕士论文

【摘要】：呼吸和心跳等生命体征反映人体的健康状况，是临床诊断和疾病预防的重要依据。常规的检测方法有接触式和非接触式两种：接触式需要依靠电极或传感器等采集信号，对人体产生约束，在一些非常规的应用场景（如烧伤、传染等）不宜使用。常用的非接触式检测方法如红外、激光等受到温度和环境障碍物的影响，应用范围受到限制。雷达式生命体征检测技术通过雷达发射电磁波照射人体，接收的回波经过信号处理分离出呼吸、心跳生命信息。鉴于该技术可以非接触、远距离、穿透障碍物，随着雷达技术、微电子技术的发展，在生命体征检测领域受到越来越多的关注。现阶段雷达式生命体征检测技术研究主要集中在对呼吸和心跳信号的检测上，，在以下方面还存在一定的技术难题：首先，由于呼吸、心跳等生命信息具有微弱、低频、易受干扰、非平稳的特点，雷达系统的系统噪声会影响回波信号的质量，所以如何提高雷达系统的稳定性、抗干扰能力成为该检测技术的难题之一。其次，呼吸运动引起的胸廓微动远大于心跳信号，且呼吸信号的谐波成分与心跳信号的频谱有重叠，所以选择怎样的信号分离方法从回波信号中分离出能量较小的心跳信号成为难题之一。针对以上问题，本文在课题组前期研究的基础上，重点研究呼吸和心跳的分离方法，主要完成以下工作： 1、分析了呼吸、心跳生理特点，研究其与体表胸廓微动的对应关系，明确了雷达检测胸廓的微动信息从中提取呼吸、心跳特征参数的生理学依据。 2、结合本研究呼吸、心跳的特点，以及自适应对消算法的原理，提出了基于LMS自适应谐波抵消的算法，通过仿真构建信号模型，用Matlab实现算法，从体动、呼吸信号中分离出心跳信号，输出结果显示仿真效果良好。 3、设计完成了相关实验，建立了接触式同步检测心电信号的对比分析实验系统，验证了算法的可行性。实验包括：实验一，模拟临床监护的人体平躺实验；实验二：模拟家庭监护的人体坐姿实验（分为深呼吸和自由呼吸两种模式）。通过采集不同实验条件下的呼吸和体动信号，并进行自适应谐波抵消处理及滤波后的输出信号与同步检测的心电信号进行对比分析。结果表明两者的频率有很强的相关性，证实了本算法可以分离出心跳信号。本课题的创新点： 1、分析了呼吸和心跳与体表胸廓微动的对应关系，明确了雷达检测呼吸心跳信号的生理学依据。 2、提出了将呼吸信号的谐波组合作为自适应滤波器的参考信号输入的基于LMS自适应谐波抵消算法，该改进算法对呼吸和心跳的分离具有较好的效果。
[Abstract]:Vital signs such as respiration and heartbeat reflect the health status of human body and are the important basis for clinical diagnosis and disease prevention. The conventional detection methods are contact-type and non-contact-type: the contact-type needs to rely on electrodes or sensors to collect signals, which has constraints on human body, and should not be used in some unconventional application scenarios (such as burn, infection, etc.). Common non-contact detection methods such as infrared laser and so on are affected by temperature and environmental obstacles and their application scope is limited. Radar vital sign detection technology irradiates the human body through radar electromagnetic wave, and the received echo signal is processed to separate the vital information of breath and heartbeat. With the development of radar technology and microelectronics technology, more and more attention has been paid in the field of vital sign detection. At present, the research of radar vital sign detection is mainly focused on the detection of respiratory and heartbeat signals, and there are still some technical problems in the following aspects: First of all, because the vital information such as breathing and heartbeat are weak, low frequency, easy to be interfered and non-stationary, the system noise of radar system will affect the quality of echo signal, so how to improve the stability of radar system, The ability of anti-jamming has become one of the difficult problems of the detection technology. Secondly, the chest fretting caused by respiratory movement is much larger than the heartbeat signal, and the harmonic components of the respiratory signal overlap with the spectrum of the heartbeat signal. So it is a difficult problem to select the signal separation method to separate the low energy heartbeat signal from the echo signal. In view of the above problems, this paper focuses on the separation method of respiration and heartbeat based on the previous study of the research group, mainly accomplishing the following work: 1. The physiological characteristics of respiration and heartbeat were analyzed, the corresponding relationship between them and the fretting of chest was studied, and the physiological basis for extracting the parameters of respiration and heartbeat from the fretting information of radar detection was established. 2. According to the characteristics of breathing and heartbeat, and the principle of adaptive cancellation algorithm, an adaptive harmonic cancellation algorithm based on LMS is proposed. The signal model is constructed by simulation, and the algorithm is realized by Matlab. The heartbeat signal is separated from the respiratory signal, and the output results show that the simulation results are good. 3. The relative experiment is completed, and the experiment system of contact-synchronous ECG signal analysis is established, which verifies the feasibility of the algorithm. The experiment includes: experiment 1, the human body lying flat experiment which simulates clinical monitoring; experiment 2: the human sitting posture experiment which simulates family monitoring (divided into two modes: deep breathing and free breathing). The respiratory and body motion signals under different experimental conditions were collected, and the output signals after adaptive harmonic cancellation and filtering were compared with those of synchronous ECG signals. The results show that there is a strong correlation between the two frequencies, and it is proved that the proposed algorithm can separate the heartbeat signal. The innovation of this topic: 1. The relationship between respiration and heartbeat and body surface chest fretting is analyzed, and the physiological basis for radar detection of respiratory heartbeat signal is clarified. 2. An adaptive harmonic cancellation algorithm based on LMS is proposed, which takes the harmonic combination of respiratory signal as the reference signal input of adaptive filter. The improved algorithm has a good effect on the separation of breath and heartbeat.
【学位授予单位】：第四军医大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R318.0

【参考文献】