Isar Mostafanezhad

Application of empirical mode decomposition in removing fidgeting interference in doppler radar life signs monitoring devices

Empirical Mode Decomposition has been shown effective in the analysis of non-stationary and non-linear signals. As an application in wireless life signs monitoring in this paper we use this method in conditioning the signals obtained from the Doppler device. Random physical movements, fidgeting, of the human subject during a measurement can fall on the same frequency of the heart or respiration rate and interfere with the measurement. It will be shown how Empirical Mode Decomposition can break the radar signal down into its components and help separate and remove the fidgeting interference.

Cancellation of Unwanted Doppler Radar Sensor Motion Using Empirical Mode Decomposition

The operation of microwave Doppler radar for sensing physiological motion signals is heavily compromised under sensor motion. To that end, we investigate the feasibility of applying empirical mode decomposition method in this context, and demonstrate its effectiveness in removing sensor motion artifacts. This method is shown to be effective in canceling unwanted sensor motion with precision sufficient to enable accurate heart rate extraction. Theoretical analysis and simulation results illustrate the potential of the proposed approach for a wide range of frequency separation and amplitude ratios of physiological signals and motion artifacts. Experimental results confirm that separation success is not very sensitive to amplitude ratio. A heart rate is extracted with RMSE within 1 beat per minute even in the presence of mechanical motion and order of magnitude larger in amplitude than that of the heart signal.

Cancellation of unwanted motion in a handheld Doppler radar used for non-contact life sign monitoring

Human life signs can be monitored at a distance using mono-static Doppler radars operating in the CW mode. Such systems rely on phase modulation of the radar’s reflected signal due to physiological motion. If the antenna is also moving, this motion will combine with and might obscure physiological motion. In a hand-held radar the antenna motion must be compensated to enable life signs extraction. It will be shown that prior knowledge of antenna’s unwanted motion can be used to overcome this problem. A signal model has been developed and experimental results confirm that it is possible to extract heart signal in the presence of antenna motion with proper compensation.

An RF Based Analog Linear Demodulator

Quadrature homodyne receivers have been used extensively for wireless life sign monitoring applications. In this letter, a new single mixer receiver architecture is proposed, that uses a phase tuning method to reach optimum demodulation point, and remove the dc offset associated with homodyne receivers. This method significantly simplifies the receiver architecture, and eliminates the demodulation step of processing baseband data. Experimental results obtained with this system demonstrate that physiological signals have been obtained without distortion, and with estimated heart rate accuracy of 99% for a stationary subject.

A coherent low IF receiver architecture for Doppler radar motion detector used in life signs monitoring

Continuous Wave Doppler radar has been used in human life signs monitoring from a distance. Such systems are basically motion detectors that rely on phase modulation of the radars reflected signal due to physiological motion which is of a low frequency nature and has significant signal content close to DC. Homodyne receiver architecture is simple, but has its own limitations including DC offset and contribution of low frequency noise from mixers and baseband amplifiers. A coherent low IF architecture has been proposed for this case to improve the performance. It will be shown that SNR is improved using a simple coherent low IF configuration. This is the first reported coherent low IF transceiver architecture for Doppler radar motion sensing.

Sensor Nodes for Doppler Radar Measurements of Life Signs

Human life signs can be detected at a distance using mono-static Doppler radars operating in the CW mode. Phase stability of the measurement system plays an important role in successful life signs detection. In this paper it will be shown that small unwanted mechanical motions of the transmit antenna, result in unrecoverable phase errors in the received signal. To overcome this issue, we propose to use a bi-static radar with a sensor node receiver placed in the vicinity of the human subject. Theoretical and experimental results confirm the benefits of using sensor nodes.

Benefits of Coherent Low-IF for Vital Signs Monitoring Using Doppler Radar

Homodyne receiver systems have been used extensively in wireless life signs monitoring applications. The main advantage of a homodyne system is range correlation resulting in cancellation of the oscillator phase noise. However, direct down-conversion to dc and the subsequent baseband amplification circuit will introduce additional flicker noise to the signal. Physiological signals have significant content around dc that will make them susceptible to 1/f noise. A coherent low-IF system is applied to solve this problem. This architecture has the range correlation benefits of the homodyne system, while minimizing the baseband flicker noise. Measurements on a mechanical target and a human subject demonstrate a signal-to-noise ratio improvement of 7 dB, which can increase the range of operation by 50%. Measurements on a human subject have demonstrated low-IF heart rate detection with a root-mean-square error of less than 0.8 beats/min at a distance of almost 3 m with transmit power of 0.1 mW, whereas direct conversion architecture output completely failed in this case.

A low cost simple RF front end using time-domain multiplexing for direction of arrival estimation of physiological signals

Doppler radar has been used for vital signs monitoring. Multiple subject monitoring and direction of arrival estimation is possible by using multiple antennas and receivers. The homodyne receiver requires a quadrature demodulator for each antenna. The complexity of the system and the number of baseband channels will increase with the number of antennas. In this paper we propose time-domain multiplexing of the RF outputs of a four element antenna array and down-conversion using a single IQ demodulator. We de-multiplex the outputs in baseband and estimate the DOA for mechanical targets and a human subject breathing in front of the radar.

Wearable sensor system for monitoring body kinematics

Existing human body motion capture solutions rely on camera based systems limited to confined measurements, or Inertial Measurement Units (IMUs) prone to noise and drift, resulting in position inaccuracies. This investigation demonstrates a proof-of-concept wearable sensor system which accurately monitors human body kinematics in real-time using Radio Frequency (RF) positioning sensors combined with MEMS based IMU sensors. In certain IMU orientations, we measured an average pitch error of <; 2 degrees for the combined method, compared with 12 degrees for an IMU alone. This self-contained sensor network has applications including military training, gaming, sports and healthcare.

Amplitude modulation issues in Doppler radar heart signal extraction

Medical Doppler radar research has largely been limited to obtaining respiratory and heart rates. While this information is vital for many applications, medical Doppler radar signatures carry significant other information that could lead to cardiopulmonary volume assessments, including cardiac stroke volume (SV), and cardiac output (CO). Accurate recovery of heart signal amplitude is required for these assessments. This paper presents the first analysis of amplitude modulation artifacts on heart signal recovery in Doppler radar systems. The sources of amplitude modulation artifacts are identified, including limitations of linear demodulation, and inherent affects of respiratory signal harmonics on heart signals. Experimental and simulation results demonstrate the validity of this analysis, and outline the path towards successful heart signal amplitude recovery.