A 5.8-GHz microwave (vital-signs) Doppler radar (for non-contact human vital-signs detection)

Transcription

1 A 5.8-GHz microwave (vital-signs) Doppler radar (for non-contact human vital-signs detection) 1 2 CB-CPW : conductor-backed coplanar waveguide At 5.8 GHz Transmission (S 21 ): -6.0 db Reflection (S 11 ): 33.5 db Isolation (S 31 ): db

2 3 4

3 2015 Student Design Competition Winners IMS High-Sensitivity Radar Giving Doppler More Bounce Chia-Hung Chao, Tzu-Wei Hsu, and Chao-Hsiung Tseng T he 5.8-GHz microwave high-sensitivity radar system we describe in this article was presented as part of the student design competitions during the 2015 IEEE Microwave Theory and Techniques Society (MTT-S) International Microwave Symposium (IMS) in Phoenix, Arizona. The goal of our competition, sponsored by the MTT and MTT-20 Technical Societies, was to design, construct, measure, and demonstrate a highsensitivity, low-power portable monostatic radar. The target to be Chia-Hung Chao (m @mail.ntust.edu.tw), Tzu-Wei Hsu (m @mail.ntust.edu.tw), and Chao-Hsiung Tseng (chtseng@ieee.org) are with the Department of Electronic and Computer Engineering, National Taiwan University of Science and Technology, Taipei, Taiwan. Digital Object Identifier /MMM Date of publication: 10 December / IEEE January 2016

4 Considering its easy realization and null-point avoidance, we chose a continuous-wave (CW) Doppler radar architecture with a quadrature receiver to detect the frequency of the vibration. tested was a moving metallic plate with a small periodic vibration. Considering its easy realization and null-point avoidance, we chose a continuous-wave (CW) Doppler radar architecture with a quadrature receiver [1], [2] to detect the frequency of the vibration. In addition, we employed only one antenna (i.e., a monostatic architecture) in the developed radar system to radiate the EM wave and then capture the scattered wave from the metallic target. The monostatic approach ensured that the main beam of the antenna would point to the target with the maximum EM energy illumination and that the maximum power would then be scattered backward from the same direction to increase the radar s sensitivity. Testing and Judging Environment istockphoto.com/dinn It was a requirement that the radar system developed for the competition be tested using a moving metallic target of 12 cm # 8 cm with periodic sinusoidal and triangular oscillations. As shown in Figure 1, the target was mounted on a linear movement stage, located 1 m away from the radar. A 1.5 m # 1.5 m reflector screen was employed to block undesired clutter noise. The actuator of the stage was programmed to have motion ranges of 10 nm, 0.1 mm, 0.5 mm, 1 mm, and 2 mm with random frequency settings of 0.2, 0.3, 0.4, 0.5, 0.6, and 0.8 Hz. At each motion frequency, the radar system had 40 s to identify the motion frequency. The figure of merit (FOM) for the radar sensor was calculated as FOM radar = ing the laptop), Tresponse is the actual time (in seconds) for detecting the 2-mm motion at 0.4 Hz. Compared with the radar-testing specification for the 2014 IMS student competition, for 2015 the detectable motion range was more stringent, and the minimum motion range, A min, was down to 10 nm. Based on (1), if the developed radar can detect the target motion, A min, at as small a distance as possible, the FOM is significantly increased. The developed 5.8-GHz radar system focused on providing the best circuit performance to detect the smallest motion range possible. To accomplish this, we chose the monostatic radar architecture shown in Figure 2 to maximize the received scattered field. Compared with a quasi-monostatic radar, as shown in Figure 2, the monostatic radar because it uses only one antenna to transmit and receive EM energy can effectively avoid the main-beam alignment loss illustrated in Figure 2. Monostatic radars usually need ferrite circulators to provide high isolation between the transmitting (Tx) and receiving (Rx) ports of the radar module. Rather than using a circulator in our design, however, we developed a passive isolation device, integrated with the radar module on a single printed circuit board (PCB) to reduce the cost and weight of the radar module. In addition, a coupler-based Tx canceller [3] and a self-injection locking radar technique [4] were employed to implement the monostatic radar shown in Figure 2. System Block Diagram The block diagram of the developed Doppler radar system is shown in Figure 3. The 5.8-GHz RF source of the radar module was provided by a Screen to Block Moving Clutter Noise Metallic Plate (1.5 m # 1.5 m) (12 cm # 8 cm) Actuactor 1 m Above Ground 1, 000, (1) Pdc # W # A min # Tresponse 1m where Pdc is the dc power consumption in mw, A min is the minimum detectable motion range (peak to peak), W is the weight of the radar sensor in grams (exclud- January 2016 Laptop Radar 1m Figure 1. The setup of the testing and judging environment. 53

5 Table 1. The manufacturer and part number of the components used in the radar module. Description VCO Driver amplifier Manufacturer and Part Number Hittite HMC431LP4 RFDM NBB400 dc Power Supply Port LNA 3.3-V LDO Driver Amplifier IF Amplifier VCO Mixer LNA IF amplifier Voltage regulator Hittite HMC488MS8G Hittite HMC320MS8G Texas Instruments LM358 Texas Instruments LM1117 z y x 2 # 2 Antenna Array 90 Phase Delay Mixer Power Divider Antenna Array Moving Target Moving Target Main Beam Alignment Loss Isolation Antenna Device Array Tx Rx Tx Rx Doppler Radar Doppler Radar Figure 4. A photograph of the developed 5.8-GHz microwave radar module. 1 (Tx) Tx Path Tx Path 100 X X (Rx) 3 (To Mixer) 70.7 X, m g /4 Rx Path 2 (Antenna) Figure 2. Comparison of a monostatic and a quasimonostatic radar architecture Arduino (ADC) I Q IF Amplifier VCO Radar Mixer LNA Isolation Device Mixer 90 Power Divider Antenna Figure 3. The block diagram of the developed 5.8-GHz microwave radar system. voltage-controlled oscillator (VCO). After this signal was amplified by a driver amplifier, the RF signal was sent to the isolation device, which consists of two Wilkinson power dividers for isolating the Tx and Rx ports of the radar module. Through the isolation device, half of the RF power is radiated into the air via the antenna. Another quarter of the transmitted signal is coupled into the receiving link for the Doppler frequency demodulation. Although one-fourth of the RF energy is dissipated by the power divider Figure 5. The schematic and the CB-CPW circuit layout of the isolation device for isolating the Tx and Rx ports. 3 resistors, using two power dividers to realize the isolation device is an effective low-cost broadband solution compared to using a hybrid coupler or a commercial isolator. As the radiating electromagnetic wave hits the moving metallic plate, the wave is scattered back with a Doppler frequency shift. The backward scattered field is received by the antenna and then amplified by a low-noise amplifier (LNA). The modulated signal is then sent into in-phase ( I ) and quadrature-phase ^Qh mixers to acquire the I- and Q-baseband signals (the Doppler frequency). Finally, an analog-to-digital conversion (ADC) module with 10 bits of ADC resolution (here an Arduino Mega 2560) was employed to convert the I- and Q-baseband signals to digital signals for further digital signal processing. 54 January 2016

6 0 0 Isolation (db) GHz 6.13 GHz -25 db Reflection Coefficient (db) db at 5.8 GHz Measured Simulated Frequency (GHz) Frequency (GHz) Figure 6. The Monte Carlo analysis of the isolation results between port 1 (the Tx port) and port 3 (the Rx port). Figure 8. The measured and simulated reflection coefficients of the two-by-two microstrip antenna array. S-Parameter (db) db at 5.8 GHz db at 5.8 GHz -30 S 11 S S db at 5.8 GHz Frequency (GHz) Figure 7. The simulated S-parameters of the CB-CPW isolation device. The radar module (including the antenna, the RF and baseband circuits, and the voltage regulation circuits) was fabricated on a RO4003 substrate with a thickness of 0.73 mm, a dielectric constant of 3.38, and a loss tangent of The active components, such as the VCO, mixer, LNA, driver amplifier, intermediate-frequency (IF) amplifier, and voltage regulator circuits were realized using the commercial integrated circuits listed in Table 1. To reduce the dielectric loss, the passive components, such as the power dividers, 90 phase delay, and connecting lines between components, were designed using conductor-back coplanar waveguide (CB-CPW). A photograph of the developed radar module is shown in Figure 4. Isolation Device Design Figure 5 shows the schematic of the isolation device adopted in the developed radar system of Figure 3. Two Wilkinson power dividers were employed to imple- Rather than using a circulator in our design, we developed a passive isolation device, integrated with the radar module on a single printed circuit board. ment this four-port isolation device to provide good isolation performance between the Tx and Rx ports. As the RF signal is launched at port 1 (the Tx port), half of the RF energy (the blue-line path) is sent directly into the successive divider and then outputs from port 2 (the antenna port), with an additional 3-dB insertion loss (IL) (namely, a 6-dB IL from port 1 to port 2). The other half of the RF power (the green-line path) couples to port 4 and is treated as the local oscillating signals for the I and Q mixers. For the receiving path, because the scattered RF signal is incident into port 2 (the antenna port), half of the received power (the red-line path) is split to port 3 (the Rx port) and then amplified by the LNA. Instead of a bulky ferrite circulator, this isolation technique is a cost-effective solution that allows the Tx and Rx ports to share only one antenna. This isolation technique can also be easily integrated with other passive and active components on a PCB. The circuit layout of the CB-CPW isolation device is shown in Figure 5, and Monte Carlo analysis using the Agilent Advanced Design System is given in Figure 6. In this analysis, 100 Monte Carlo samples were used to randomly vary the port impedance of the isolation device from 45 X to 55 X (50 X!10%). The Monte Carlo analysis shows the isolation is better than 25 db from 5.67 GHz to 6.13 GHz with a relative bandwidth of 7.8%. Figure 7 shows the simulated S-parameters of the developed isolation January

7 Radar dc Power Supply Computer for Signal Processing Arduino for Data Acqusition -120 Copolarization Cross-Polarization m Moving Metallic Plate (12 cm # 8 cm) Figure 10. The complete 5.8-GHz radar system and the testing environment device. At 5.8 GHz, the transmission ^S21h, reflection ^S11h, and isolation ( S31 ) coefficients are -6.0 db, db, and db, respectively. Antenna Design A two-by-two microstrip antenna array is used to transmit and receive the radar system s RF signals (Figure 4). The initial size of the single antenna element was calculated using Ansoft Designer at 5.8 GHz, and an inset feeding technique was then employed to match the impedance to 50 X. To improve the side-lobe level (SLL) of the antenna array, the center-to-center spacing between the two adjacent antenna elements is 074. m 0, 38 mm at 5.8 GHz. The one-to-four feed network was implemented using T junctions and quarterwave transformers. Figure 8 shows the measured and simulated reflection coefficients of the developed antenna array. The 0 Copolarization Cross-Polarization Figure 9. The measured E-plane (yz-plane) and H-plane (xz-plane) radiation patterns of the two-by-two microstrip antenna array at 5.8 GHz. I and Q Channel Voltage (V) Normalized Spectrum I Channel Q Channel Time (s) Frequency (Hz) Figure 11. The measurement results of the time-domain waveforms and the normalized frequency spectrum as the metallic plate moves with a 100-nm variation at a frequency of 0.2 Hz. 56 January 2016

8 antenna array s frequency range (assuming a reflection coefficient of better than db) is from 5.77 GHz to 5.86 GHz with a 1.72% relative bandwidth. Because the developed radar system only works at a single frequency, 5.8 GHz, the impedance bandwidth of the designed antenna array is acceptable for the CW Doppler application. The measured E-plane and H-plane radiation patterns of the antenna array at 5.8 GHz are illustrated in Figure 9. For the E-plane pattern [Figure 9], the antenna gain was dbi, the 3-dB beam width was 15, and the cross-polarization level (CPL) was 22 db. For the H-plane pattern [Figure 9], the antenna exhibited an antenna gain of dbi, a 3-dB beam width of 14.5, and a CPL of 26 db. The measured radiation efficiency was 50.63% at 5.8 GHz, and the SLL of the antenna array was better than 10 db. Signal Processing The outputs of the radar module (i.e., the demodulated I- and Q-baseband signals) are sent to an Arduino Mega 2560 module to perform the ADC, and then acquired by National Instruments Lab- VIEW software (Figure 3). The complex signal demodulation method [5] is used to combine the I and Q signals as a complex number. By applying the fast-fourier transformation (FFT), the resulting frequency spectrum can be used to determine the vibration frequency of the metallic plate. Measurements The complete 5.8-GHz radar system is shown in Figure 10. The radar module is powered by a voltage of 5 V and a current of 223 ma, with a total dc power ^Pdch of 1115 mw. The weight ^Wh of the radar and Arduino modules, including the connecting USB cables, was 123 g. To evaluate the developed radar system in our laboratory, as shown in Figure 10, a metallic plate of 12 cm # 8 cm was located 1 m away from the radar, and mounted on the linear motor stage to provide a periodical 100-nm vibration at a frequency of 0.2 Hz. Using a lapse of 14 s ^Tresponseh, the measured time-domain waveforms of the I and Q channels become stable as shown in Figure 11. Figure 11 shows that a measurement distance of 1 m makes the Q channel approach the null point of the radar system, but the I channel approaches the optimum point. By applying the FFT, the acquired time-domain waveforms were transformed into the frequency domain in real time, as shown as Figure 11. The peak frequency is located at 0.2 Hz, the same as the frequency of the actuator. Due to the wavelength of the transmitted EM wave, the minimum detectable motion range ^Aminh of the developed 5.8-GHz radar To receive the maximum scattered field from the moving metallic target, the Tx and Rx ports of the radar module share a single antenna array using a PCB-based isolation device to realize a monostatic radar architecture. was 100 nm. Therefore, the FOM of the radar in (1) was determined as FOM radar = Conclusions This article has described the development of a 5.8-GHz microwave Doppler radar. The radar system s detection capability has been demonstrated under the testing environment diagrammed in Figure 1. To receive the maximum scattered field from the moving metallic target, the Tx and Rx ports of the radar module share a single antenna array using a PCB-based isolation device to realize a monostatic radar architecture. Due to the radar s operating frequency of 5.8 GHz, the minimum detectable motion range of the developed radar is limited to 100 nm. In the future, a higher operating frequency, such as 24 GHz, could be chosen to obtain a smaller detectable motion range. Acknowledgment This work was supported by the Ministry of Science and Technology of Taiwan under Grant E The authors also thank the National Chip Implementation Center of Hsinchu, Taiwan, for the PCB fabrication. References [1] A. D. Droitcour, O. Boric-Lubecke, V. M. Lubecke, J. Lin, and G. T. A. Kovacs, Range correlation and I/Q performance benefits in sigle-chip silicon Doppler radars for noncontact cardiopulmonary monitoring, IEEE Trans. Microwave Theory Tech., vol. 52, no. 3, pp , Mar [2] C. Li, V. M. Lubecke, O. Boric-Lubecke, and J. Lin, A review on recent advances in Doppler radar sensors for noncontact healthcare monitoring, IEEE Trans. Microwave Theory Tech., vol. 61, no. 5, pp , May [3] C.-Y. Kim, J.-G. Kim, and S. Hong, A quadrature radar topology with Tx leakage canceller for 24-GHz radar applications, IEEE Trans. Microwave Theory Tech., vol. 55, no. 7, pp , July [4] F.-K. Wang, T.-S. Horng, K.-C. Peng, J.-K. Jau, J.-Y. Li, and C.- C. Chen, Single-antenna Doppler radars using self and mutual injection locking for vital sign detection with random body movement cancellation, IEEE Trans. Microwave Theory Tech., vol. 59, no. 12, pp , Dec [5] C. Li and J. Lin, Random body movement cancellation in Doppler radar vital sign detection, IEEE Trans. Microwave Theory Tech., vol. 56, no. 12, pp , Dec January