PORTABLE NON-CONTACT MICROWAVE DOPPLER RADAR FOR RESPIRATION AND HEARTBEAT SENSING

Transcription

1 Armenian Journal of Physics, 2012, vol. 5, issue 1, pp PORTABLE NON-CONTACT MICROWAVE DOPPLER RADAR FOR RESPIRATION AND HEARTBEAT SENSING H. Avagyan, A. Hakhoumian, H. Hayrapetyan, N. Pogosyan, T.Zakaryan Institute of Radiophysics and Electronics, National Academy Sciences of Armenia, 1 Alikhanian St., Ashtarak, 0203, Armenia Received 13 April, 2012 Abstract Microwave L-band Doppler radar has been developed for pulmonic and cardiac activities investigation. Model of reflected signal spectra has been proposed, which is in qualitative agreement with both pulmonic and cardiac spectra. Distance sensitivity and penetration ability of the radar allows to quality it for typical indoor applications, as well as for life signal detection under avalanche. 1. Introduction Microwave Doppler radar has been used for wireless sensing applications for many years. Beginning from 1970s, microwave Doppler radar found new applications in human healthcare monitoring and detection. It offers new opportunities, such as physiological movement and volume change sensing [1], human vital signal detection for finding trapped people under earthquake rubble [2]. First works were done with heavy and bulky waveguides, but recent advances in microwave and radar technologies made it possible to integrate such a system on a single chip [3,4], which is compact, light-weight and low-cost. With inexpensive and compact design, microwave Doppler radar could be used in home healthcare monitoring, particularly for detecting sleep apnea [5,6]. Microwave Doppler radar was first used for sensing of respiration rate and the detection of apnea in 1975 [1]. Since 1980s, similar systems were developed for finding victims trapped in earthquake rubble and an avalanche [7] and sensing human presence behind a wall or other barriers [8]. All these systems were designed using bulky and heavy microwave components and large antennas, which are acceptable for using in diagnostic institutions, but are impractical for home healthcare monitoring. Alternatives to this for heart and respiration home monitoring are polar straps [9], chest expansion measuring straps [10] for respiration monitoring, acoustic monitors, nasal and oral sensors. All these methods require contact with body and careful placement. Compact Doppler radar may provide portable and more flexible noncontact alternative. Doppler-type motion-sensing radar systems typically transmit continuous-wave (CW) signal (sometimes frequency-modulated), which is reflected off the target and received by the receiver. According to Doppler theory, a moving target will cause frequency shift in transmitted signal, which can be detected by detector. A stationary person has chest movement, and, therefore, Doppler radar with chest as target will receive transmitted signal with Doppler shift caused by movement of the chest, which contains information about heartbeat and respiration. By using existing wireless and radar technologies, it would be possible to design inexpensive and portable device performing such monitoring.

2 Portable Non-Contact Microwave Doppler Radar Armenian Journal of Physics, 2012, vol. 5, issue 1 In this paper we propose to use separate receive and transmit antennas in order to provide more isolation vs. circulator case and to investigate human heartbeat and respiration. 2. Overall Design The block-diagram of the proposed radar is schematically presented in Fig. 1. The system operates at 1 GHz operating frequency generated by Mini-Circuits ZX C local oscillator (LO). The signal is amplified and fed to microstrip patch transmit antenna. Received signal via another identical antenna is down-converted to zero-frequency baseband and amplified by video amplifier to satisfy the ADC dynamic range requirements. Fig. 1. Block-diagram of the radar Antenna System, Noise and Bandwidth Consideration Using of microwave technology for the detection of cardiac and pulmonic activity mainly is caused by the possibility of considerable penetration of electromagnetic waves into the barrier. This, on the one hand, requires to use relatively low frequencies (hundreds of megahertz). However, it is obvious that on the other hand there is a serious problem of the radiation wave length due to unacceptably bulky and traditional antenna systems. Using microstrip fractal patch antennas can simultaneously solve a number of problems. The two most important ones are small size and low cost. Specially designed antennas have geometric dimensions shown in Fig. 2, and have beamwidth of about 30º (Fig. 3). The sensitivity of a microwave direct conversion Doppler radar for vital sign detection depends significantly not only on the strength the respiration and cardiac signals but also the effect of noise at both RF and baseband frequencies. Since received signals are small, they are very sensitive to the noise. Therefore, it is important to understand the sources of noise in the system. There are two main sources of noise in radar system affecting detection of physiological signals: thermal noise and Flicker noise. These two noise sources are taken into account separately at RF frequency and then combined at baseband after the mixer. 9

3 H. Avagyan et al. Armenian Journal of Physics, 2012, vol. 5, issue 1 Fig. 2. Geometry of microstrip fractal patch antenna. Fig. 3. Directivity plot for microstrip patch antenna in H- (red) and V-(blue) planes. Then, it is well known that the devices of biological purpose need to use amplifiers with low noise figure and high thermal stability. Proper restoration of the heartbeat needs an amplifier working within the bandwidth of Hz. For the normal heart (60 beats per minute) low frequency limit will start with a 1 Hz, while for the faulty one (40 beats per minute) its value is reduced to 0.16 Hz and even less. Upper frequency limit depends on the typical speed of heart valves which is between 0.05 to 0.1 seconds. Respective frequency is Hz. Considering the fact that the mixer output signal has DC component, which inevitably occupies the dynamic range, the amplifier for alternate subsonic signals has been chosen, which passes with unity-gain the DC component. The lower limit of such a filter should be 5 10 times smaller than the lowest operating frequency, i.e. fcutoff Hz Hz. Because the working frequency range lies in the band with significant flicker noise, we chose AD8675 OPAMP, which has a low noise spectral density of 2.8 nv Hz Reflected Waveform Let us consider the reflected signal spectrum from oscillating target. We present transmitter signal in the form 10

4 Portable Non-Contact Microwave Doppler Radar Armenian Journal of Physics, 2012, vol. 5, issue 1 Tr t, i tr Ae (1) where the oscillation phase tr t t (2) represents the phase of the harmonic oscillator. Respectively, for the reflected and received signal will have the form Rx t, i r Be x (3) where rx t D t t (4) and the amplitude and phase in (4) are determined by many parameters such as target RCS, target range, etc. Doppler frequency shift is defined as [14] And finally for the signal phase we have and for the entire reflected received signal () t 2 v() t 2v cos Ω t t. (5) D rx 2 cosω m m t t v t t (6) it i 2vm cos Ω t t x. (7) R Be e Thus, the spectrum of the signal reflected from oscillating target is identical to the spectrum of FM signal. Such consideration can explain the periodic-like spectrum of pulmonic and cardiac activities. Here modulation index m is determined by the amplitude of oscillator 2 v. Spectrum of FM signal with modulation index m less than 0.5 represents itself main tone and only two sidelobes with opposite phases. While for the modulation indexes m greater than 1, there are many spectral components (theoretically infinite numbers), but we can neglect the members of Fourier series with nm 2 [13]. Characteristic features corresponding to overtone structure can be used for diagnostic purposes by entering the metric 2 m2 2 Bk B 0 R 1, (8) k 1 Jk ( m) J0( m) where J ( m ) is the Bessel function of first kind k-th order from real argument. k m 3. Measurements and discussions We have investigated various human cardiac and pulmonic activities. The spectral resolution of the radar is determined by observation time (~20 60 sec) and can distinguish the pulmonic activity from the cardiac one. The most intensive is the reflection from the respiratory organs due to their relatively high RCS. 11

5 H. Avagyan et al. Armenian Journal of Physics, 2012, vol. 5, issue 1 In order to achieve better understanding, we carried out separate investigations of cardiac activity and mixed pulmonic/cardiac activities. Figs. 4 and 5 show typical cardiac and mixed pulmonic/cardiac activities. It is remarkable, that there are only a few sidelobs in pure cardiac spectra, so we can assign a small modulation index ( m 0.5) to that activity [13]. In contrast, we can t say the same for pulmonic activity where many sidelobes are presented. Therefore pulmonic activity can be described by relatively large modulation indexes ( m 1) [13]. Fig. 4. Typical mixed cardiac/pulmonic activity. Upper time domain; lower power spectrum of received signal strength (RSS). Fig. 5. Typical cardiac activity. Upper time domain; lower power spectrum of received signal strength (RSS). 12

6 Portable Non-Contact Microwave Doppler Radar Armenian Journal of Physics, 2012, vol. 5, issue 1 Distance sensitivity of non-contact radar is shown in Fig. 6. It is obvious that such a system can effectively used in typical indoor applications with operational radius up to a few meters. Fig. 6. Reflection intensity vs. distance. Another important feature of microwave radar is penetration ability. Fig. 7 shows how much can penetrate radiation before it achieves the human body. We chose as a barrier the concrete blocks with various thicknesses. Fig. 7. Reflection intensity vs. barrier thickness. 4. Conclusion The designed radar can be used as a tool for studying human respiratory and cardiac activity, as well as life sensor for earthquakes rubbles. Spectral resolution satisfies to distinguish cardiac activity from respiratory one, and hence makes it possible to investigate spectral peculiarities for diagnostic purposes. 13

7 H. Avagyan et al. Armenian Journal of Physics, 2012, vol. 5, issue 1 REFERENCES 1. D.Samardzija, B.-K.Park, O.Boric- Lubecke., V.Lubecke, A.Host-Madsen, T.Sizer, Experimental Evaluation of Multiple Antenna Techniques for Remote Sensing of Physiological Motion, Microwave Symposium, 2007, IEEE/MTT-S International, p K.M.Chen, Y.Huang, J.Zhang, A.Norman, IEEE Trans. Biomedical Engineering, 27, 105 (2000). 3. A.D.Droitcour, O.Boric- Lubecke, V.Lubecke, J.Lin, G.T.A. Kovacs, Int. Solid-State Circuits Conf. Dig., 1, San Francisco, CA, 2002, p A.D.Droitcour, O.Boric- Lubecke, V.Lubecke, J. Lin, G.T.A. Kovacs, IEEE MTT-S Int. Microwave Symp. Dig., 3, Philadelphia, PA, 2003, p M.D.Bethesda, Infantile apnea and home monitoring, NIH, Consensus Statement, 6, September D.Droitcour, O.Boric- Lubecke, G.T.A.Kovacs, IEEE Transactions on Microwave Theory and Techniques, October 2009, p K.M.Chen, Y.Huang, J.Shang, A.Norman, IEEE Trans. Biomed. Eng., 27, 105 (2000). 8. J.Seals, S.R.Crowgey, S.M.Sharpe, Georgia Tech. Res. Inst., Atlanta, GA, Final Rep. Project A , R.M.T.Laukkanen, P.K.Virtanen, J. Sports Sci., 16, S3 (1998). 10. K.Leino, S.Nunes, P.Valta, J.Takala, Acta Anaesthesiologica Scandinavica, 45(1), 104 (2001). 11. A.Hakhoumian, H.Hayrapetyan, S. Martirosyan, A.Muz hikyan, N.Pogh osyan, T.Z akaryan, The International Conference of The Technique of Microwave and THz Waves and its Application in Biomedical and Radar Technologies and in Remote Sensing, (IRPhE 2010), September 23 25, 2010, Ashtarak-Aghveran, Armenia, p A.Muzhikyan, A.Hakhoum ian, S.Martirosyan, V.Nikoghosyan, N.Poghosyan, T.Poghosyan, K.Rustamyan, T.Z akaryan, Proc. of 11-th International Radar Symposium IRS-2010, June 16 18, 2010, Vilnius, Lithuania, p D.Calcutt, L.Tetley, Satellite Communications Principles and Applications, Butterworth-Heinemann, N.Levanon, E.Mozeson, Radar Signals, John Wiley & Sons,