A Miniature Mode Reconfigurable Inductorless IR-UWB Transmitter–Receiver … ·...

294 IEEE JOURNAL ON EMERGING AND SELECTED TOPICS IN CIRCUITS AND SYSTEMS, VOL. 8, NO. 2, JUNE 2018

A Miniature Mode Reconfigurable InductorlessIR-UWB Transmitter–Receiver for Wireless

Short-Range Communicationand Vital-Sign Sensing

Zhe Zhang , Member, IEEE, Yongfu Li, Member, IEEE, Koen Mouthaan, Member, IEEE,and Yong Lian, Fellow, IEEE

Abstract— This paper presents a miniature inductorlessimpulse-radio ultra-wideband transmitter–receiver and radarfor wireless short-range communication and vital-sign sensing.The all-digital transmitter generates impulse-radio ultra-wideband pulses using edge combining technique and con-sumes 21.6 pJ/pulse at 10 Mb/s. A novel active-inductor-basedtechnique is proposed for low noise amplifier that achievesultra-wideband input impedance matching in the receiver. Thenon-coherent receiver employs a simple demodulation andsynchronization circuit to achieve self-synchronization withoutany on-chip or external oscillator. Consuming a total powerof 6.4 mW, the receiver attains −64-dBm sensitivity at 10 Mb/s.The chip is implemented in a 65-nm digital CMOS processand occupies only 0.04 mm2. Measurement shows that thetransmitter–receiver and radar is capable of sensing the human’srespiratory rate when the time interval is measured by a fastsampling digital oscilloscope.

Index Terms— Impulse-radio, ultra-wideband, all-digitalIR-UWB transmitter, inductorless, self-synchronization, contact-less sensing, non-coherent, ultra-wideband radar, vital signdetection, active inductor.

I. INTRODUCTION

THERE are increasing interests in low-power wirelesstechnologies for biomedical applications. Wireless data

communication is the most common application in wirelessbody area network, as shown in Fig. 1(a). Another applicationis contact-less vital sign detection such as radar for remotesensing of heartbeat and respiration as shown in Fig. 1(b). Forall these applications, low emission level, low power, and lowcost are essential requirements. Thus, a configurable chip that

Manuscript received September 15, 2017; revised December 12, 2017;accepted January 15, 2018. Date of publication January 30, 2018; dateof current version June 11, 2018. This work was supported by NSERCDiscovery Grant. This paper was recommended by Guest Editor P.-I. Mak.(Corresponding author: Zhe Zhang.)

Z. Zhang was with the Department of Electrical and Computer Engineering,National University of Singapore, Singapore 117576. He is now with HuaweiInternational Pte. Ltd., Singapore 117674 (e-mail: [email protected]).

Y. Li and K. Mouthaan are with the Department of Electrical and ComputerEngineering, National University of Singapore, Singapore 117576 (e-mail:[email protected]; [email protected]).

Y. Lian is with the Department of Electrical Engineering and Com-puter Science, Lassonde School of Engineering, York University, Toronto,ON M3J 1P3, Canada (e-mail: [email protected]).

Color versions of one or more of the figures in this paper are availableonline at http://ieeexplore.ieee.org.

Digital Object Identifier 10.1109/JETCAS.2018.2799930

Fig. 1. Wireless techniques in biomedical applications (a) Wireless bodyarea network and wireless sensor network (b) Contact-less sensing of vitalsigns.

serves both wireless data communication and remote sensingis highly welcomed in wireless biomedical applications.

Among all the wireless techniques for biomedicalapplications, impulse-radio ultra-wideband (IR-UWB) is ofspecial interest. The emission level allowed by FederalCommunications Commission (FCC) for IR-UWB signal(−41.3 dBm/MHz) is very low and the potential harmfulphysiological effect caused by the electromagnetic radiation

2156-3357 © 2018 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.

https://orcid.org/0000-0002-4839-8770

ZHANG et al.: MINIATURE MODE RECONFIGURABLE INDUCTORLESS IR-UWB TRANSMITTER–RECEIVER 295

is weak [1]. The large bandwidth of the IR-UWB signaltranslates to very narrow pulses in the time domain. The pulse-like nature of the IR-UWB signal not only enables heavy duty-cycling of the circuit which reduces the power consumption,but also allows high-resolution localization and sensing. Thispaper presents a complementary metal-oxide-semiconductor(CMOS) IR-UWB transmitter-receiver which can be recon-figurable into both short-range wireless communication modeand vital-sign sensing mode. This work also addresses therequirements of IR-UWB systems for biomedical applicationsincluding power and cost reductions.

Low power consumption is the foremost objective in wire-less wearable sensors. The energy efficiency of the IR-UWBtransmitter has been greatly improved by techniques suchas aggressive duty-cycling and all-digital implementa-tion [2]–[4]. The challenges lie in the receiver. Comparedwith the coherent receiver [5], [6], the non-coherent IR-UWBreceiver reduces the circuit complexity and power consumptionof the radio front-end [7]–[11]. However, the simplicity ofthe non-coherent IR-UWB front-end is penalized by the com-plex demodulation and synchronization [12]. Synchronized-OOK (S-OOK) modulation was proposed in [7] to achievelow power timing synchronization and data reception. In [7],an on-chip ring oscillator generates the receiver clock, whosefrequency is fixed and prone to process variations. In this work,we eliminate the use of the on-chip or off-chip clock generatorand the data is recovered by full self-synchronization. Thesimple demodulator and synchronizer are composed of onlydigital gates, which helps achieve low power consumption andhigh energy efficiency.

Continuous and accurate monitoring of human vital signsis important for telemedicine and homecare. Such monitoringcan be achieved by either wearable devices or contactless vitalsign sensing. Contact-less vital sign sensing has an advan-tage over wearable devices as it is non-intrusive. Single-chipnarrow-band Doppler radar for vital sign detection has beenreported with high resolution [13]. CMOS IR-UWB radarswere reported in [14]–[19]. Correlation-based receiver [15]and sampling-based receiver [14], [16] require large chiparea and high power consumption. Threshold sampler basedradar utilizes continuous-time signal processing in [17] or1-b direct RF sampling in [19]. In [18], three-dimensionalbeamforming is achieved by a 16-channel (4×4) planar timedarray with large delay range and fine delay resolution. Allthe existing IR-UWB radar transceivers are designed specifi-cally for radar applications, and most of the reported CMOSIR-UWB receivers extensively use large-area monolithicinductors to improve the input matching and enhance the gainbandwidth [8], [10]. In this paper, we explore the use of sameIR-UWB non-coherent transmitter-receiver for both communi-cation and radar applications. We employ inductorless designsfor both transmitter and receiver for reduced area. Real-timesensing of human’s respiration is achieved.

This paper is organized as follows. Section II intro-duces the system architecture and calculates the link budget.Section III and IV discuss the design considerations and circuitimplementations. Section V presents the measurement resultsand conclusion is given in Section VI.

Fig. 2. System architecture of IR-UWB system (a) Short-rangecommunication (b) Radar sensor.

II. SYSTEM ARCHITECTURE AND LINK BUDGET

The proposed IR-UWB system is configurable for bothshort-range communication and radar sensing. The systemarchitecture is shown in Fig. 2, where Fig. 2(a) representsthe communication mode and Fig. 2(b) represents the radarsensing mode. All-digital implementation is adapted for trans-mitter design to improve the energy efficiency. An active-inductor-based low noise amplifier (LNA) is proposed in thereceiver for better wideband matching and chip area reduction.Non-coherent pulse detector downconverts the IR-UWB pulsesto baseband. The difference between the communication modeand radar mode is in the modulator and demodulator. In thecommunication mode, S-OOK modulation and demodulationare performed. In the radar mode, there is no modulationand a train of IR-UWB pulses with a fixed repetition rateis transmitted and received. This is realized simply by settingall the input symbols as ’0’s in the S-OOK modulator. Thetime interval between the transmitted and received pulses ismeasured to determine the distance between the radar sensorand the detected object.

The link budget calculation for the short-range IR-UWBsignal transmission is summarized in Table I. The IR-UWBsignal is pulse-based and the link budget can be calcu-lated in terms of transmitted and received pulse energy [9].IR-UWB signal covers a bandwidth of several hundreds ofmegahertz or several gigahertz. It is difficult to calculate thepulse energy by integrating the power spectral density acrossthe ultra-wide signal band since the the power spectral densityat each specific frequency could not be accurately modelled.The transmitted pulse energy is thus estimated by circuitsimulation using the equation as below:

E pT X = VppT X2

8RLwp Np (1)


Fig. 3. Schematic of the IR-UWB transmitter.

TABLE I

LINK BUDGET CALCULATION

VppT X is the simulated TX output peak-to-peak voltage, RL isthe output impedance, wp is the width of one elementary pulse,Np is the number of elementary pulses in one TX pulse. In thecommunication mode, the transmitter-receiver is to supportthe link between the wireless sensors and the Central ControlUnit (CCU) as shown in Fig. 1(a). The transmission distance is1 m or less. The path loss is calculated as 50 dB according tothe channel model in [20]. To achieve a total error rate of 10−3

in the S-OOK modulation, it is shown that the required Eb/N0is 18 dB [21]. The noise figure of the non-coherent receiveris as high as 20 dB due to the receiver’s non-linear nature [9].A link margin of 3 dB can be obtained from the calculation.In the radar sensor mode, the transmitter-receiver is placednot far from the human body as shown in Fig. 1(b). Large-size antennas with directional gain of 10 dB are used. In thisconfiguration, the signal can be received at a longer distancethan in communication mode. Only the link budget calculationfor communication mode is shown in Table I. The calculationpredicts that the non-coherent receiver with a relatively simplesystem architecture is capable of supporting short-rangeIR-UWB signal transmission and reception.

III. TRANSMITTER DESIGN

Both analog and digital approaches have been appliedin generating the IR-UWB pulses. The analog methods ofgenerating pulses use up-conversion [22] or digital-to-analogconverter direct synthesis [23]. However, the high powerconsumption of both approaches is not suitable for low-powerbiomedical applications. Compared with the analog counter-parts, all-digital IR-UWB transmitter is built from digitalgates which consumes low power and provides good energyefficiency. One digital approach that avoids up-conversion isto use an on-chip bandpass filter to form a narrow pulse [24].The shortcoming of such approach is large chip area, whichincreases the cost. The IR-UWB transmitter in this workexploits direct edge combining technique to synthesize theoutput waveform [3], [22]. There is no on-chip inductor sothe chip area is minimized.

The circuit of the transmitter is shown in Fig. 3. The S-OOKwaveform is produced by the modulator. One pilot pulse andone data pulse are generated by the modulator if the input non-return-to-zero (NRZ) data is ’1’. Only the pilot pulse but nodata pulse is present if the data is ’0’. The modulator is com-posed of all-digital gates and the data pulse lags the pilot pulseby 5 ns. The modulated pulses then propagate into the variabledelay line which produces rising and falling edges with smalldelays between them. The delay line is composed of current-starved asymmetric inverters whose delays are controlled bya 10-bit digital word. The asymmetric inverter architecturesuppresses the undesirable glitches and improves the rise/falltime [25]. The edges are then combined by logic AND and ORgates to generate the negative and positive impulses. Six delaycells are used as the delay line and the number of oscillationsin the pulse is four. Four impulses are chosen to cover thefrequency band from 3 GHz to 8 GHz. 16-bit amplitude tuning(4 bit for each impulse) controls the number of active ANDand OR gates. The impulses drive the digital PAs and the dualcapacitively-coupled structure with precharge transistors helps


to reduce the low frequency components [2], [3]. Theprecharge transistors are self-controlled by the first and lastpropagating edges V0m , V6m , V0p , V6p using simple combina-tion logic gates.

The power consumption is divided into active power andleakage power. When the data rate is low, leakage powerconsumed by the digital gates significantly degrades the powerefficiency. Optimal sizing of the transistors starts from theoutput drivers. The transistors of the output drivers are sized todrive the antenna with a 0.8 V output voltage swing. Then theAND and OR gates are sized followed by the interstage buffersand the delay cells. Transistors at each stage are sized to drivethe later stages with acceptable delay and rise/fall time. Themaximum repetition frequency of the pulses is 80 MHz.

On-chip decoupling capacitors are added to reduce thesupply voltage ripple. The parasitics of bondpads, ESD devicesand bondwires are modelled and included in the simulation.In this design, the transmitter generates a UWB signal thatcovers 3.1-8 GHz. The simulated pulse peak-to-peak amplitudeis about 0.8 V at a supply voltage of 1.2 V. The peak-to-peakoutput amplitude is limited by the 1.2 V power supply voltageand output transistor size. Large output transistors add a largeload to the AND and OR gates and make driving the outputtransistors at high frequency difficult.

IV. RECEIVER DESIGN

The IR-UWB receiver includes an active-inductor-basedwideband LNA, a pulse detector and a demodulator andsynchronizer.

A. Active-Inductor-Based Self-Biased Wideband LNAVarious techniques have been proposed to achieve wide-

band input matching and gain in LNAs. The common-gate (CG) LNA was widely used due to its superiorbroadband input impedance matching and better linearityperformance [26]. In the CG LNAs, the input capacitancedegrades the input impedance matching significantly whenthe operating frequency increases to several gigahertz. Thetotal input capacitance includes the gate-source capacitance ofthe input transistor and other parasitic capacitance from thepackage and ESD devices. An inductor is often introduced toform a resonant network and neutralize the input capacitance.In this work, an active-inductor-based input impedance match-ing technique for CG LNA is intoduced. The on-chip passiveinductor is replaced by an active inductor which is composedof M1 and a feedback resistor RF . The output of the firststage of the LNA is connected to the gate of M1 through aresistor RF , as shown in Fig. 4. Circuit analysis shows thatthe input admittance can be expressed as:

Gin = gm1gm2 RL11/( jωCgs1)

1/( jωCgs1) + RF+ jωCt + gm2

= gm1gm2 RL1

1 + ω2Cgs1 RF+gm2− jω

Cgs1gm1gm2 RL1 RF

1 + ω2Cgs1 RF+ jωCt

(2)

gm1 and gm2 are the transconductance of M1 and M2. Cgs1 isthe gate-source capacitance of M1. RL1 is the load of thefirst stage. Ct is the total input capacitance which degrades

Fig. 4. Schematic of the active-inductor-based wideband LNA.

the input matching in traditional common-source (CS) or CGamplifiers. It is shown in (2) that the active inductor intro-duces a negative susceptance which cancels the positive sus-ceptance Ct . The second stage of the LNA is a cascodeCS stage for gain boosting. The third stage is a CS stagewith an active-inductor load (by Rg and Cby) for bandwidthextension [27], [28]. The multi-stage LNA topology providessufficient gain and good input matching over very widebandwidth. The LNA is designed to work with a 1 V powersupply and is self-biased without the need of biasing circuit.

The noise factor of the active-inductor-based LNA is esti-mated as below [29], [30].

F = 1 + γ2

gm2 RS+ γ1gm1 RS + (RL1 + RF )

RS(1 + gm2 RL1)2 (3)

RS is the 50 � source impedance. γ1 and γ2 are the excessivenoise coefficients of M1 and M2. The second term in (3) isintroduced by the transistor M2. The third term is introducedby the transistor M1 in the active-inductor. The fourth termis introduced by the two resistors, RL1 and RF . Note that alarger RF results in larger negative suspectance but increasednoise factor. The value of the RF is chosen in such a way thatthe input matching bandwidth and noise factor are balanced.

The procedure of designing the first stage of LNA includestwo steps. First a CG stage is designed to match the inputimpedance to source at lower frequency, then RF is tuned tocancel the impact of input capacitance on input matching athigher frequency. Fig. 5 shows the simulation results of theinput matching |S11| and gain |S21| across different processcorners. Fig. 6 compares the simulated input matching, gainand NF with and without active-inductor feedback. The active-inductor feedback helps achieve wide-band input matchingand gain. Active-inductor feedback generates more noise. Thenoise figure (NF) of 5 to 7 dB is acceptable for short-range communication or radar sensing with the link budgetcalculated in Section II. The wide-band input matching isachieved from 3.1 GHz to 8 GHz. The simulated voltagegain is from 23 to 25 dB and the simulated NF is from5 to 7 dB within the band of interest. The gain of 23 to 25 dBis sufficient to amplify the input voltage to about 50 mV,which can trigger the pulse detector in the following stage.The power consumption of the LNA is 5.8 mW. The transistor


Fig. 5. Simulated |S11| and |S21| of LNA in different process corners(a) |S11|, (b) |S21|.

TABLE II

LNA DEVICE SIZING

sizes and the passive component values of the LNA are listedin Table. II.

B. Pulse DetectorIn traditional non-coherent IR-UWB receivers, a self-

mixer or squarer is used to downconvert the UWBsignal [7], [9]. The baseband signal is then amplified bya baseband gain amplifier or integrated by an integrator.An analog-to-digital converter or a comparator generates thedigital output. In this work, the pulse detector is significantlysimplified as shown in Fig. 7. A common-source amplifier,utilizing the non-linear relationship between the drain currentand the gate voltage, was used as a pulse detector [31], [32].The resistor load is replaced with a PMOS transistor to achievea high conversion gain. The input NMOS transistor is biasedby Vn in the subthreshold region and consumes little powerwhen no pulse is received. PMOS transistor is biased by Vp sothat the output of the CS amplifier is about the threshold levelof the inverter. The nonlinear I-V relationship of the NMOStransistor generates the baseband signal as shown in Fig. 7(b).

Fig. 6. Simulated |S11| |S21| and |N F | of LNA, with feedback (wFB) andwithout feedback (woFB). (a) |S11|, (b) |S21|, (c) |N F |.

Fig. 7. Schematic and mechanism of the pulse detector.

Three stages of asymmetrically-sized inverters amplifies thebaseband signal to full scale. The sizes of the inverter chaintransistors are shown in Fig. 7. When the PMOS is much larger


than NMOS, the inverter pulls up the rectified pulse hard.When the NMOS is much larger than PMOS, the inverter pullsdown the rectified pulse hard. The buffer then further rectifiesthe baseband signal and produces the digital output pulses.The pulse detector is AC coupled with the LNA. The inverterstages are DC coupled to reduce the number of bias voltagesrequired. The asymmetrically-sized inverters amplifies thepulse to a large amplitude so that AC coupling is not required.The pulse detector consumes 80 μW when the data rateis 10 Mbps.

C. Demodulation and SynchronizationIn S-OOK, one symbol is composed of two pulses, namely

the pilot pulse and the data pulse. The pilot pulse indi-cates the start of one symbol and the data pulse carries thedata. The circuit of demodulator and synchronizer is shownin Fig. 8 (a) and the timing diagram of the demodulation andself-synchronization is shown in Fig. 8 (b).

The baseband pulse generated by the pulse detector is toonarrow to trigger the D flip-flop (DFF) in the demodulator.Thus, we use two pulse stretching circuits (PSCs) to expendthe pulses from the pulse detector. PSC_A expands both thepilot pulse and the data pulse to the width of T1. PSC_B onlyexpands the pilot pulse since the stretched pulse width T2 islarger than the delay between the pilot pulse and data pulse.The output of the PSC_B is used as the reference clock. Theoutput of PSC_A is latched by this reference clock using aDFF. T2 is tuned by four binary bits (B0-B3) so that the fallingedge of the reference clock can latch the stretched data pulseif the symbol is ’1’. When the symbol is ’0’, there is no datapulse and the output of the DFF is low.

The output data is in NRZ format. Only 13 logic elements(3 flip-flops and 10 combinational gates) are used in thedemodulator and synchronizer. This is much simpler than [21]which includes 61 logic elements (16 flip-flops and 45 combi-national gates). No on-chip or off-chip clock source is requiredfor the receiver.

V. MEASUREMENT RESULTS

The test chip is fabricated in a 65 nm digital CMOS 1P6Mprocess. The chip is packaged in standard QFN40 package andmounted on Rogers 4003 PCB for testing (Fig. 19).

A. TransmitterThe measured time-domain transmitter output is shown

in Fig. 9(a). The output peak-to-peak voltage amplitude is0.75 V. The power consumption is 216 μW at 10 Mbps andthe figure of merit (FOM) of energy per bit is 21.6 pJ/bit. Thespectrum of the transmitted pulse is shown in Fig. 9(b) and iscompliant with FCC regulations. The transmitter can supportup to 80 Mbps data transmission.

B. Active-Inductor-Based Wideband LNAThe simulated and measured input |S11| are shown

in Fig. 10. The measured |S11| is below -9.5 dB from3.1-7.8 GHz. The |S11| is measured on the packaged chipwhich is mounted on PCB. Good input matching is achievedacross very wide band despite of the parasitic effects from

Fig. 8. (a) Schematic of the demodulator (b) Timing diagram of improvedS-OOK demodulation and synchronization.

bondpads, package, routing trace and connector. This showsthe effectiveness of the proposed active-inductor-based match-ing technique. The DC current consumption of the LNA is6.3 mA from a 1 V power supply.

C. Short-Range Communication ModeFig. 11(a) shows the block diagram of the measurement

setup for the communication mode. A variable attenuatoris used for the sensitivity test. The free space LOS test isshown in Fig. 11(b). A pair of omni-directional antennas areused [33]. The transmitted and received data are captured bya Tektronix DPO 71254 oscilloscope and shown in Fig. 12.Note that the output data is in NRZ format which can bedirectly processed by a baseband processor. The plot of BERversus the receiver input power is shown in Fig. 13. The inputsensitivity is -64 dBm for a bit-error-rate (BER) of 10−3 whenthe data rate is 10 Mbps.

The transmitted pulse energies are similar in measurementresults and link budget because of the similar peak-to-peakpulse amplitude and impulse width. The measured free-air


Fig. 9. Transmitter output in (a) time domain (b) frequency domain.

Fig. 10. Simulated and measured |S11|.

communication distance for BER of 10−3 is about 0.7 m,which can be interpreted as 3 dB worse sensitivity comparedto the link budget. The measured sensitivity is -64 dBm atdata rate of 10 Mbps. This indicates minimum received pulse

Fig. 11. Measurement setup for communication mode. (a) Block diagramof the test setup (b) LOS free space test environment.

Fig. 12. Transmitted and received NRZ data.

energy as 40 aJ, which is also 3 dB higher than that in thelink budget.

D. Radar Mode - Respiratory Rate MeasurementThe setup for measuring the human respiration by the

radar mode is shown in Fig. 14. The distance between theantenna and body is about 40 cm. The pulse rate is 1 Mbpswhile the chest movement rate is under 10 Hz. The chest isapproximately still for many pulses and the movement can bereconstructed by time of flight (ToF) estimation. Directional


Fig. 13. BER versus input power.

Fig. 14. Measurement setup for radar sensor mode. (a) Block diagram ofthe test setup (b) Test environment for measuring human respiration.

antennas with 10 dB gain are used to increase the transmissiongain [34]. The antennas are placed in parallel so the interfer-ence from the transmitting antenna to the receiving antennais minimized. The digital output of the receiver is sampledand recorded by a Tektronix DPO 71254 oscilloscope. Theframe sample rate is 3 Hz and the time resolution is 4 ps. TheToF estimation is based on the equation D = C × T/2. D isthe distance between the radar and the object. C is the speedof the pulse which is estimated as a constant. T is the timeinterval for the pulse travelling between the radar and object.

Fig. 15. Respiration measurement (a) normal respiration in time domain(b) normal respiration in frequency domain (c) reference from commercialmonitoring module.

Note that the time interval can be obtained by an on-chiptime-to-digital converter. The time interval is then convertedto the displacement by a computer program. The displacementof the chest versus time during normal and fast respirationis shown in Fig. 15(a) and Fig. 16(a). Fourier transforma-tion reveals the respiratory rate as shown in Fig. 15(b) andFig. 16(b). The respiration frequency is 0.25 Hz (15 rpm) fornormal respiration and 0.7 Hz (42 rpm) for fast respiration.In order to verify the accuracy of the measurement results,a commercial hospital-standard Philips IntelliVue X2 mod-ule is used to measure the respiration rate simultaneouslyas the reference. The measured respiration rates from thereference module are 14 rpm for the normal respiration and46 rpm for the fast respiration. The results from the IR-UWB


Fig. 16. Respiration measurement (a) fast respiration in time domain (b) fastrespiration in frequency domain (c) reference from commercial monitoringmodule.

transmitter-and-receiver in radar mode matches the resultsfrom the reference equipment.

Fig. 17 shows one case of real-time human respirationmonitoring. For the first 45 seconds, normal respiration isdetected. The breath is then halted at the 45th second andis resumed after 20 seconds. Since the breath was held for20 seconds, the respiratory rate is faster than normal whenresumed.

E. Performance Comparison

Table III shows the performance of the presented transmitterand several previous publications. Three FOMs are compared.The first FOM is the widely-used energy per pulse EdT X ,which is the total power consumption divided by the pulse rate.

Fig. 17. Human respiration changes from normal respiration to heldrespiration to resumed respiration.

Fig. 18. Micrograph of the chip.

Fig. 19. Test boards for (a) Transmitter (b) Receiver.

The second FOM is the transmission efficiency, which isdefined as the ratio of transmitted pulse energy E pT X toEdT X [9]. The efficiency is important since the transmittedpulse energy is closely related to the communication rangein non-coherent IR-UWB transceiver. However, the E pT X isdifficult to measure and is usually estimated from the timedomain waveform or the power spectrum of the pulse. Thetransmission efficiency of the IR-UWB transmitter is usuallybelow 10% [3], [9]. The third FOM is the energy per pulsenormalized to output voltage amplitude EdT X/Vpp , whichwas introduced in [24]. The energy per pulse EdT X of this


TABLE III

TRANSMITTER PERFORMANCE COMPARISON

TABLE IV

RECEIVER PERFORMANCE COMPARISON

TABLE V

CMOS VITAL-SIGN RADAR PERFORMANCE COMPARISON

work is 21.6 pJ/pulse, which is better than or comparable tostate-of-the-art works. The presented transmitter features anefficiency of 7.5%, which is the same as [3] and higher thanthe others. The pulse energy normalized to the output voltageamplitude is 27 pJ/V, which is equal to that in [24]. Note thatthere are multiple on-chip inductors in [24] and the area isvery large. The transmitter in this work only occupies an areaof 0.03 mm2, as shown from the die photo in Fig. 18.

Table IV compares the designed IR-UWB receiver withother designs. Because of the simple demodulation and syn-chronization circuit, our receiver achieves an energy per bit

of 0.64 nJ/bit, which is lower than other references. Thesensitivity of the receiver is measured as -64 dBm at 10 Mbps,which allows a reliable wireless data link for short distance.The total area of the designed receiver is only 0.01 mm2,which includes both the radio front-end and the demodulator(Fig. 18). Note that the works in [3] and [35] achieves very lowpower and energy per bit through duty-cycling. This techniquecould be explored to reduce the power consumption of thiswork.

Table V compares the vital-sign sensing performance of thiswork with other CMOS designs. This work achieves the lowest


power consumption and smallest die area at the expenses ofshorter detection range.

VI. CONCLUSION

In this paper, a fully integrated 65 nm CMOS IR-UWBtransmitter-receiver and radar for short-range communicationand human respiration monitoring is presented. The trans-mitter generates output with 0.75 V peak-to-peak amplitude,consuming 21.6 pJ/pulse at 10 Mbps. The receiver achievesself-synchronization with a simple demodulation and synchro-nization circuit. When the data rate is 10 Mbps, the receiverachieves -64 dBm sensitivity and consumes 6.4 mW. Thetransmitter-receiver achieves the best FOMs compared withstate-of-the-art references. Free-air short-range communicationis tested and the feasibility of applying the non-coherenttransmitter-receiver in real-time human respiration sensingis verified. The transmitter-receiver is fully inductorless andexhibits low area, low circuit complexity and low powerconsumption.

REFERENCES

[1] First Report and Order, Revision of Part 15 of the Commissions RulesRegarding Ultra-Wideband Transmission Systems, document ET Docket98-153, FCC, Washington, DC, USA, Apr. 2002.

[2] P. P. Mercier, D. C. Daly, and A. P. Chandrakasan, “An energy-efficient all-digital UWB transmitter employing dual capacitively-coupled pulse-shaping drivers,” IEEE J. Solid-State Circuits, vol. 44,no. 6, pp. 1679–1688, Jun. 2009.

[3] L. Wang, Y. Lian, and C.-H. Heng, “3–5 GHz 4-channel UWBbeamforming transmitter with 1° scanning resolution through calibratedVernier delay line in 0.13-μm CMOS,” IEEE J. Solid-State Circuits,vol. 47, no. 12, pp. 3145–3159, Dec. 2012.

[4] S. Diao, Y. Zheng, and C. H. Heng, “A CMOS ultra low-power andhighly efficient UWB-IR transmitter for WPAN applications,” IEEETrans. Circuits Syst. II, Exp. Briefs, vol. 56, no. 3, pp. 200–204,Mar. 2009.

[5] S. Joo et al., “A fully integrated 802.15.4a IR-UWB Transceiver in0.13 μm CMOS with digital RRC synthesis,” in IEEE ISSCC Dig. Tech.Papers, Feb. 2010, pp. 228–229.

[6] F. Zhang, A. Jha, R. Gharpurey, and P. Kinget, “An agile, ultra-widebandpulse radio transceiver with discrete-time wideband-IF,” IEEE J. Solid-State Circuits, vol. 44, no. 5, pp. 1336–1351, May 2009.

[7] M. Crepaldi, C. Li, J. R. Fernandes, and P. R. Kinget, “An ultra-wideband impulse-radio transceiver chipset using synchronized-OOK modulation,” IEEE J. Solid-State Circuits, vol. 46, no. 10,pp. 2284–2299, Oct. 2011.

[8] Y. J. Zheng et al., “A 0.92/5.3nJ/b UWB impulse radio SoC forcommunication and localization,” in IEEE ISSCC Dig. Tech. Papers,Feb. 2010, pp. 230–231.

[9] S. Solda, M. Caruso, A. Bevilacqua, A. Gerosa, D. Vogrig, andA. Neviani, “A 5 Mb/s UWB-IR transceiver front-end for wireless sensornetworks in 0.13-μm CMOS,” IEEE J. Solid-State Circuits, vol. 46,no. 7, pp. 1636–1647, Jul. 2011.

[10] D. Lachartre et al., “A 1.1nJ/b 802.15.4a-compliant fully integratedUWB transceiver in 0.13 μm CMOS,” in IEEE ISSCC Dig. Tech. Papers,Feb. 2009, pp. 312–313.

[11] D. C. Daly et al., “A pulsed UWB receiver SoC for insect motioncontrol,” IEEE J. Solid-State Circuits, vol. 45, no. 1, pp. 153–166,Jan. 2010.

[12] P. P. Mercier, M. Bhardwaj, D. C. Daly, and A. P. Chandrakasan, “A low-voltage energy-sampling IR-UWB digital baseband employing quadraticcorrelation,” IEEE J. Solid-State Circuits, vol. 45, no. 6, pp. 1209–1219,Jun. 2010.

[13] C. Li, X. Yu, C.-M. Lee, D. Li, L. Ran, and J. Lin, “High-sensitivitysoftware-configurable 5.8-GHz radar sensor receiver chip in 0.13- μmCMOS for noncontact vital sign detection,” IEEE Trans. Microw. TheoryTechn., vol. 58, no. 5, pp. 1410–1419, May 2010.

[14] T.-S. Chu, J. Roderick, S. Chang, T. Mercer, C. Du, and H. Hashemi,“A short-range UWB impulse-radio CMOS sensor for human featuredetection,” in IEEE ISSCC Dig. Tech. Papers, Feb. 2011, pp. 294–296.

[15] D. Zito et al., “SoC CMOS UWB pulse radar sensor for contactlessrespiratory rate monitoring,” IEEE Trans. Biomed. Circuits Syst., vol. 5,no. 6, pp. 503–510, Dec. 2011.

[16] P. Park, S. Kim, S. Woo, and C. Kim, “A centimeter resolution,10 m range CMOS impulse radio radar for human motion monitor-ing,” IEEE J. Solid-State Circuits, vol. 49, no. 5, pp. 1125–1134,May 2014.

[17] H. A. Hjortland, D. T. Wisland, T. S. Lande, C. Limbodal, and K. Meisal,“Thresholded samplers for UWB impulse radar,” in Proc. IEEE Int.Symp. Circuits Syst., May 2007, pp. 1210–1213.

[18] L. Wang, K. H. Teng, Y. Lian, and C. H. Heng, “A 4×4 IR UWB timed-array radar based on 16-channel transmitter and sampling capacitorreused receiver,” IEEE Trans. Circuits Syst. II, Exp. Briefs, to bepublished.

[19] N. Andersen et al., “A 118-mW 23.3-GS/s dual-band 7.3-GHz and8.7-GHz impulse-based direct RF sampling radar SoC in 55-nm CMOS,”in IEEE ISSCC Dig. Tech. Papers, Feb. 2017, pp. 138–139.

[20] Channel Model for Body Area Network—IEEE 802.15 Working GroupDocument. Accessed: Feb. 1, 2018. [Online]. Available: https://math.nist.gov/mcsd/savg/papers/15-08-0780-09-0006-tg6-channel-model.pdf

[21] M. Crepaldi and P. R. Kinget, “Error ratio model for synchronised-OOKIR-UWB receivers in AWGN channels,” Electron. Lett., vol. 49, no. 1,pp. 25–27, Jan. 2013.

[22] D. D. Wentzloff and A. P. Chandrakasan, “A 47pJ/pulse 3.1-to-5GHz all-digital UWB transmitter in 90nm CMOS,” in IEEE ISSCC Dig. Tech.Papers, Feb. 2007, pp. 118–591.

[23] D. Baranauskas and D. Zelenin, “A 0.36 W 6 b up to 20 GS/s DAC forUWB wave formation,” in IEEE ISSCC Dig. Tech. Papers, Feb. 2006,pp. 2380–2389.

[24] S. Bourdel et al., “A 9-pJ/pulse 1.42-Vpp OOK CMOS UWB pulsegenerator for the 3.1–10.6-GHz FCC band,” IEEE Trans. Microw. TheoryTechn., vol. 58, no. 1, pp. 65–73, Jan. 2010.

[25] C. Hu and P. Y. Chiang, “All-digital 3-50 GHz ultra-wideband pulsegenerator for short-range wireless interconnect in 40 nm CMOS,” inProc. IEEE Custom Integr. Circuits Conf., Sep. 2011, pp. 1–4.

[26] H. Zhang, X. Fan, and E. S. Sinencio, “A low-power, linearized, ultra-wideband LNA design technique,” IEEE J. Solid-State Circuits, vol. 44,no. 2, pp. 320–330, Feb. 2009.

[27] B. Razavi, Design of Integrated Circuits for Optical Communications.New York, NY, USA: McGraw-Hill, 2002.

[28] S. K. Hampel, O. Schmitz, M. Tiebout, and I. Rolfes, “Inductorless1–10.5 GHz wideband LNA for multistandard applications,” in Proc.IEEE Asian Solid-State Circuits Conf., Nov. 2009, pp. 269–272.

[29] B. Razavi, “Cognitive radio design challenges and techniques,” IEEEJ. Solid-State Circuits, vol. 45, no. 8, pp. 1542–1553, Aug. 2010.

[30] J. Borremans, P. Wambacq, C. Soens, Y. Rolain, and M. Kuijk,“Low-area active-feedback low-noise amplifier design in scaled digitalCMOS,” IEEE J. Solid-State Circuits, vol. 43, no. 11, pp. 2422–2433,Nov. 2008.

[31] J. K. Brown, K.-K. Huang, E. Ansari, R. R. Rogel, Y. Lee, andD. D. Wentzloff, “An ultra-low-power 9.8 GHz crystal-less UWBtransceiver with digital baseband integrated in 0.18 μm BiCMOS,” inIEEE ISSCC Dig. Tech. Papers, Feb. 2013, pp. 442–443.

[32] A. Oncu and M. Fujishima, “19.2 mW 2 Gbps CMOS pulse receiverfor 60 GHz band wireless communication,” in Proc. IEEE Symp. VLSICircuits, Jun. 2008, pp. 158–159.

[33] Z. N. Chen, T. S. P. See, and X. Qing, “Small printed ultrawidebandantenna with reduced ground plane effect,” IEEE Trans. AntennasPropag., vol. 55, no. 2, pp. 383–388, Feb. 2007.

[34] Z. N. Chen, M. J. Ammann, X. Qing, X. H. Wu, T. S. See, andA. Cai, “Planar antennas,” IEEE Microw. Mag., vol. 7, no. 6, pp. 63–73,Dec. 2006.

[35] R. K. Dokania, X. Y. Wang, S. G. Tallur, and A. B. Apsel, “A low powerimpulse radio design for body-area-networks,” IEEE Trans. CircuitsSyst. I, Reg. Papers, vol. 58, no. 7, pp. 1458–1469, Jul. 2011.

[36] J. Ryckaert, G. V. D. Plas, V. D. Heyn, C. Desset, B. V. Poucke, andJ. Craninckx, “A 0.65-to-1.4 nJ/burst 3-to-10 GHz UWB all-digital TXin 90 nm CMOS for IEEE 802.15.4a,” IEEE J. Solid-State Circuits,vol. 42, no. 12, pp. 2860–2869, Dec. 2007.

[37] S. Gambini, J. Crossley, E. Alon, and J. M. Rabaey, “A fully inte-grated, 290 pJ/bit UWB dual-mode transceiver for cm-range wirelessinterconnects,” IEEE J. Solid-State Circuits, vol. 47, no. 3, pp. 586–598,Mar. 2012.

[38] B. Vigraham and P. Kinget, “A self-duty-cycled and synchronized UWBreceiver SoC consuming 375 pJ/b for −76.5 dBm sensitivity at 2 Mb/s,”in IEEE ISSCC Dig. Tech. Papers, Feb. 2013, pp. 444–445.


Zhe Zhang (S’09–M’16) received the B.S. degreein physics from Peking University, Beijing, China,in 2008, and the Ph.D. degree from the Departmentof Electrical and Computer Engineering, NationalUniversity of Singapore, Singapore, in 2015.

He is currently with Huawei InternationalPte. Ltd., Singapore. His research interests includeintegrated wireless circuits, integrated power con-verters, and analog-to-digital converters.

Yongfu Li (S’09–M’15) received the B.Eng. andPh.D. degrees from the Department of Electricaland Computing Engineering, National Universityof Singapore (NUS), Singapore, in 2009 and 2014,respectively.

He was a Research Engineer with NUS, from2013 to 2014. Since 2014, he has been with GLOB-ALFOUNDRIES as a design-to-manufacturing(DFM) computer-aided design (CAD) DevelopmentEngineer. His research interests include analog/mixed signal circuits, data converters, powerconverters, and DFM CAD.

Koen Mouthaan received the M.Sc. and Ph.D. degrees in electrical engi-neering from the Delft University of Technology, The Netherlands, the MBAdegree from Nanyang Technological University, the M.Sc. degree in organi-zational leadership from Johns Hopkins University, and the master’s degree inspace engineering from the Technical University of Berlin. He was with TNODefense, Safety and Security, The Netherlands, and with SkyGate, a companythat designed phased-array antennas for consumer applications. From 2003 to2015, he was an Assistant Professor and an Associate Professor with theDepartment of Electrical and Computer Engineering, National Universityof Singapore, where he rejoined in 2016. His research interests includemicrowave and millimeter-wave circuits and systems, phased array antennas,digital beamforming, design, and innovation.

Yong Lian (M’90–SM’99–F’09) received the B.Sc.degree from the College of Economics and Man-agement, Shanghai Jiao Tong University, Shanghai,China, in 1984, and the Ph.D. degree from theDepartment of Electrical Engineering, National Uni-versity of Singapore (NUS), Singapore, in 1994.He was with the industry for nine years and joinedNUS in 1996, where he was the Deputy DepartmentChair of research, the Area Director of IC andembedded systems with the Department of Electricaland Computer Engineering, a member of the Univer-

sity Tenure and Promotion Committee and Senate Delegacy. He was appointedas the first Provost’s Chair Professor with the Department of Electrical andComputer Engineering, NUS, in 2011. He is currently a Professor in YorkUniversity. His research interests include biomedical circuits and systems andsignal processing.

Dr. Lian is a fellow of the Academy of Engineering Singapore.He received many awards, including the IEEE Circuits and Systems Society’sGuillemin-Cauer Award (1996), the IEEE Communications Society Multime-dia Communications Best Paper Award (2008), the Institution of EngineersSingapore Prestigious Engineering Achievement Award (2011), the Hua YuanAssociation/Tan Kah Kee International Society Outstanding ContributionAward (2013), the Chen-Ning Franklin Yang Award in Science and Technol-ogy for New Immigrant (2014), and the Design Contest Award in 20th Inter-national Symposium on Low Power Electronics and Design (ISLPED2015).As an Educator, he received the University Annual Teaching Excellent Awardin two consecutive academic years from 2008 to 2010 and many other teachingawards from the Faculty of Engineering of NUS. Under his guidance, hisstudents received many awards, including the Best Student Paper Award inICME 2007, the winner of 47th DAC/ISSCC Student Design Contest in 2010,the Best Design Award in A-SSCC 2013 Student Design Contest.

Dr. Lian is the President of the IEEE Circuits and Systems (CAS) Society,a member of IEEE Fellow Committee, a Steering Committee Memberof the IEEE TRANSACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS.He was the Editor-in-Chief of the IEEE TRANSACTIONS ON CIRCUITS ANDSYSTEMS PART II: EXPRESS BRIEFS for two terms from 2010 to 2013.He was the Guest Editor for eight special issues in the IEEE TRANSACTIONS

ON CIRCUITS AND SYSTEMS-PART I: REGULAR PAPERS, the IEEE TRANS-ACTIONS ON BIOMEDICAL CIRCUITS AND SYSTEMS, and the Journal ofCircuits, Systems Signal Processing. He was the Vice President of Publicationsof the IEEE CAS Society from 2013 to 2015, the Vice President of the Asia–Pacific Region of the IEEE CAS Society from 2007 to 2008, an AdCommMember of the IEEE Biometrics Council from 2008 to 2009, a CAS SocietyRepresentative to the BioTechnology Council from 2007 to 2009, a Chairof the BioCAS Technical Committee of the IEEE CAS Society from 2007 to2009, a Chair of the DSP Technical Committee of the IEEE CAS Society from2010 to 2011, a member of the IEEE Medal for Innovations in HealthcareTechnology Committee from 2010 to 2012, and a Distinguished Lecturerof the IEEE CAS Society from 2004 to 2005. He is the Founder of theInternational Conference on Green Circuits and Systems, the Asia PacificConference on Postgraduate Research in Microelectronics and Electronics,and the IEEE Biomedical Circuits and Systems Conference.

A Miniature Mode Reconfigurable Inductorless IR-UWB Transmitter–Receiver … ·...

Documents

Transcript of A Miniature Mode Reconfigurable Inductorless IR-UWB Transmitter–Receiver … ·...