Multi-Port Front-End and DSP
Co-Design for Vital Signs Detector
Adriana Serban (Craciunescu), Oscar Morales, Tobias Petersson, Henrik
Kalvér and Qin-Zhong Ye
Conference article, Oral presentation
Cite this conference article as:
Serban, A., Morales, O., Petersson, T., Kalvér, H., Ye, Q-Z. Multi-Port Front-End and
DSP Co-Design for Vital Signs Detector. Oral presentation at the Swedish Microwave
Days, GigaHertz Symposium, Lund, Sweden, May 24-25, 2018.
Copyright: The Authors
The self-archived postprint version of this conference article is available at Linköping
University Institutional Repository (DiVA):
Multi-Port Front-End and DSP Co-Design for Vital
Adriana Serban1, Oscar Morales2, Tobias Pettersson3, Henrik Kalvér3, Qin-Zhong Ye1
1ITN-Communication Electronics, LiU, Campus Norrkoping, 601 74 Norrkoping 2Student, Electrical Engineering Master Programme, LiU
3Student, Electronic Design Programme, LiU
Direct conversion radio applications, such as off-chip, broadband, linear and low-power quadrature modulators and demodulators using multi-port technique were inten-sively studied at Linköping University (LiU). Controlled wave interferometry within the passive multi-port (six-port) correlator with capability to accurately process phase infor-mation is also useful in microwave radar and wireless sen-sor applications . In this paper, the co-design and imple-mentation of a complete hardware-software Doppler radar modular system for vital signs detection using the multi-port technology is presented . In contrast to multi-port radio applications, the challenge is to demonstrate the possibility to detect weak, Hz-range frequency signals.
In Fig. 1, the principle of heartbeat and breathe rate detec-tion using continuous wave (CW) Doppler radar is illus-trated. The return signal, modulated by periodical vital signs is received and further processed by the receiver front-end (Rx) acting in conjunction with the digital signal processing (DSP) block.
A. Multi-Port Front-End Unit
The multi-port front-end unit was designed to operate at 5.8 GHz (ISM band). The antennas, the channel and the entire multi-port demodulator were modelled, designed and man-ufactured. In Fig. 2, the complete model of the front-end is shown. Thermal noise, flicker noise, and phase noise were considered during the design process due to their impact on the sensor sensitivity. For hardware-software co-design, a communication path between the multi-port microwave unit model in ADS and the DSP unit was established.
A. Digital Signal Processing Unit
A dedicated DSP unit was developed so that I- and Q raw data coming from the multi-port front-end could be used by the DSP unit to verify and optimize the algorithms and demonstrate the detectability of the vital signs. The imple-mented DSP algorithms include low-pass filtering, decima-tion, arctangent calculadecima-tion, Fast Fourier transform and peak detection, as shown in Fig. 3.
In this project, we have investigated the possibility of de-tection of low power signals at very low frequencies that emulate vital signs using the multi-port technology. The multi-port front-end was designed and first prototypes were manufactured on ROGERS4350B substrate. An appropriate digital signal processing unit was developed and imple-mented to acquire and further process the raw data coming from the front-end. It was shown that, with the existing measurement equipment, 10-Hz periodical signals could be detected.
Fig. 1. Vital signs detector system, conceptual block diagram.
Fig. 2. Multi-port front-end model in ADS.
Fig. 3. Implemented digital signal processing flow.
 S. O. Tatu, A. Serban, M. Helaoui, A. Koelpin, ”Multiport Technology”, IEEE Microwave Magazine, vol. 15, no. 7, pp. S34-S44, 2014.
 Adriana Serban, Qin-Zhong Ye, “Radio and Radar Platform for Enhancing Education and Research at LiU, Campus Norrköping”, Norrköpings Fond for Forsking och Utveckling Project, 2015-2017. Tx
Antennas and Channel
MW Multi-Port Correlator and Power Detectors
Acquisition and Digital Signal Processing
Expected Results Rx Co -d es ig n Pre-processing (Filtering Decimation) ADC ADC I Q
Acquisition Mai-loop of the DSP-kernel
DMA Memory Processing (arctan FFT) Post-processing (Peak detection) Transmitting (via Ethernet or USB) User Displaying data (on a computer/server, GUI,
every 0.5 s) Interrupt Ethernet/USB-link Pre-processing (Filtering Decimation)