----Project Description----

Project objective in brief – Non-contact measurement of heartbeat and respiratory activity at close distance using stationary FMCW radar.

Motivation and importance of the project – Contactless health monitoring is convenient and can be used in particularly sensitive cases, especially in people with burns, in newborns, or psychiatric patients. When the impact of radio waves on the human body is thoroughly studied, health-monitoring radar can be one of the important paths for the development of disease-preventive systems and medical care systems.

Objectives:

1. Registration of respiratory activity and heart rate by directing a radar beam at the human body and visualizing radar signatures of vital signs. Proposed solution: In order to record the signal, the radar beam will be directed at the chest of a person located several dozen centimeters away from the antenna. For vital signs visualization, various processing methods will be used, in particular time frequency analysis and phase demodulation. MTI or CLEAN processing will be used to remove clutter if necessary. Processing algorithms and visualization of results will be implemented in MATLAB.

2. Implementation of algorithms enabling detection of individual respiratory and heart cycles and estimation of their frequency of occurrence. Proposed solution: Respiration rate can be determined quite easily, because the chest displacement signal is sinusoidal. A simple FFT should work for this purpose. Heart rate estimation is more difficult because the signal consists of high-frequency oscillations. Detection of oscillations related to the heart can be achieved in different ways. One of them is to use time-frequency methods to represent the heart oscillations in a spectrogram, which has properties that smooth high-frequency oscillations and allows them to simulate a sinusoid. Another approach could be Doppler frequency envelope analysis or convolution with matched filter model of the heart cycle. The algorithms will be developed and tested in MATLAB. The best one will be selected for use in objective 4.

3. Conducting experiments including:

   • Testing the measurement at different chest-to-antenna distances. This will allow to check the distance capabilities of the device.
   • Measurement of vital signs from different sides of the body, in particular from the front, from the back side and measurement of the heartbeat when the beam is directed at the carotid artery. The experiment will indicate the best measurement perspective for specific applications and the possibility of using alternative perspectives.
   • Simultaneous measurement of vital signs for several people located in slightly different positions relative to the radar. In particular, at different distances from the antenna and at similar azimuth, and at different azimuths and the same distance from the antenna. The aim of the experiment is to test the distance resolution due to the FMCW signal capabilities and the angular resolution due to beam steering.
   • Measurement of vital signs in a person wearing clothes of different thickness and without clothes. The experiment will allow to assess the influence of such obstacles on the measurement, in order to indicate limitations for some applications.
   • Measurement during various states of body activity, especially after exercise, in a relaxed state and during sleep. Additionally, heart activity will be measured during held breathing to obtain a reference signal. The experiment will show how the intensity of vital signs affects the signal-to-noise ratio and the ability to detect heart and respiratory cycles.
   • Attempt to measure a human running in place/on a treadmill or performing other activities such as typing on a laptop or talking. The experiment will show the limitations associated with random body movements introducing noise in vital signs signatures.

4. Implementation and testing of a radar system for real-time monitoring of vital signs. Proposed Solution: The best algorithm from objective 2 will be selected for continuous data processing. Processing will be implemented directly on the Raspberry Pi. If the computing power is insufficient, the pre-processed data will be sent to another computing platform. The results will be saved to a memory card or displayed in real time. To test the system, long-term measurements of breathing rate and heart rate will be made and compared with the results of conventional sensors such as a smartwatch or pulse oximeter.