This is the accepted version of the article: Herrojo, Cristian; Muela, Francisco Javier; Mata Contreras, Francisco Javier; [et al.]. «High-density microwave encoders for motion control and near-field chipless- RFID». IEEE sensors journal, Vol. 19, issue 10 (May 2019), p. 3673-3682. DOI 10.1109/JSEN.2019.2895015 This version is avaible at https://ddd.uab.cat/record/221406 under the terms of the license 1 High-Density Microwave Encoders for Motion Control and Near-Field Chipless-RFID Cristian Herrojo, Member, IEEE, Francisco Muela, Javier Mata-Contreras, Ferran Paredes, Member, IEEE, and Ferran Martín, Fellow, IEEE Abstract— A novel microwave system for measuring linear Encoder motion displacements and velocities with sub-millimeter resolution and for the implementation of near-field chipless radiofrequency Coupling Modulation identification (chipless-RFID) systems with very high data capacity is presented. The system is based on a reader, consisting of a half-wavelength straight resonator coupled (through Modulating signal capacitor gaps) to a pair of access lines, and a microwave encoder, in relative motion to the reader and consisting of a linear chain of strips orthogonally oriented to the chain axis. By RF Carrier AM modulator displacing the encoder over the half-wavelength resonator of the (Interrogation (Transmission AM Demodulator reader, with the encoder strips parallel oriented to the reader signal) line) axis, the relative velocity and position between the encoder and the reader can be inferred. For that purpose, the reader is fed by a harmonic signal tuned to the resonance frequency that results when an encoder strip is perfectly aligned with the reader. Encoder motion amplitude modulates the feeding signal at the Envelope signal output port, and both the position and the velocity are measured (contains the ID code, from the peaks, or dips, of the resulting envelope function. velocity or displacement) Moreover, by making certain strips inoperative, the system can be used for coding purposes. Due to the small period of the Fig. 1. General block diagram of the AM system used for measuring angular encoder (0.6 mm), a high per-unit-length data density in these or linear velocities and displacements, or for the implementation of near-field near-field chipless-RFID tags (i.e., 16.66 bits/cm) is achieved. To chipless-RFID systems. illustrate the functionality and potential of the approach, 100-bit chipless-RFID tags with various ID codes are implemented and variation in the resonance frequency that results when the read. Index Terms— Displacement sensors, velocity sensors, transmission line and the resonant element are in relative chipless-RFID, microwave encoders, motion control. motion (either linear or angular) [2],[6]. In other displacement and velocity sensors, the working principle is coupling I. INTRODUCTION modulation [3]-[5],[7]. Although frequency variation and coupling modulation sensors are simple from a technological T here are several approaches for the implementation of displacement and velocity sensors based on microwave technologies. In particular, a significant research effort has viewpoint, their resolution and dynamic range are typically limited. Moreover, frequency variation sensors may be subjected to cross sensitivities caused by environmental been dedicated in the last decade to the development of factors (e.g., frequency drifts caused by temperature variations sensing systems where a transmission line is loaded with or changes in ambient moisture), and coupling modulation resonant elements in relative motion respect to it [1]-[10] sensors are very sensitive to noise. (other sensors based on resonator-loaded lines, useful for In [8], inspired by the optical encoders, a novel approach for dielectric characterization and related applications, have been the measurement of angular displacements and velocities also reported [11]-[25]). In several of the previous spatial based on microwaves was proposed. The working principle in sensors, the relative displacement is determined from the such sensors, similar to optical encoders, was pulse detection. However, rather than apertures, which provide the relative This work was supported by MINECO-Spain (projects TEC2016-75650-R displacement and velocity between the moving object and a and RTC-2017-6303-7), Generalitat de Catalunya (project 2017SGR-1159), fixed optical beam used to illuminate the apertures in an ICREA, and by FEDER funds. C. Herrojo acknowledges MINECO for optical system, in the sensor reported in [8] the pulses were supporting his research activity through the FPI grant BES-2014-068164. generated from a chain of resonant elements, particularly split C. Herrojo, F. Muela, F. Paredes and F. Martín are with GEMMA/CIMITEC, Departament d’Enginyeria Electrònica, Universitat ring resonators [26]. In [8] and in subsequent works [9],[10], Autònoma de Barcelona, 08193 Bellaterra, Spain. E-mail: such resonator chain was etched in the edge of the rotor, and Ferran.Martin@uab.es. as many pulses per revolution as resonant elements in the J. Mata-Contreras is with the Departamento de Ingeniería de Comunicaciones, Escuela Técnica Superior de Ingeniería de chain (or twice this value) were achieved. Up to 1.200 pulses, Telecomunicación, Universidad de Málaga, 29071 Málaga, Spain. 2 0 f0 (a) l1 -10 w g -20 |S21| (dB) (b) -30 Loaded reader with operative strip l2 -40 Bare reader -50 Loaded reader with shortened inoperative strip Loaded reader with inoperative strip (cut in the middle) -60 3.8 3.9 4.0 4.1 4.2 Fig. 2. Layout of the reader (a), and three-dimensional view of the reader with Frequency (GHz) a strip on top of it and perfectly aligned (b). Dimensions are (in mm): l1= 21.2, Fig. 3. Simulated frequency responses (including losses) of the bare reader, l2= 22.4, g = 0.2 and w = 0.4. The considered reader substrate is the Rogers and reader loaded with perfectly aligned strips under different conditions, RO4003C with dielectric constant r = 3.55, thickness h = 0.81 mm and loss including operative strip with nominal length, inoperative strip with cut in the tangent tan = 0.0021. The considered encoder substrate is the Rogers middle position, and inoperative strip with shortened length. The considered RO4003C with dielectric constant r = 3.55, thickness h = 0.204 mm and loss air gap, or vertical distance between the tag and the reader is 0.2 mm. In all tangent tan = 0.0021. the simulations a pair of adjacent operative strips has been considered. providing an angular resolution of 0.3º, were demonstrated in [9]). For pulse generation, caused by rotor motion, a in this paper) is fundamental in order to implement high data transmission line conveniently designed, acting as the stator of capacity systems compatible with short tags (the 100-bit tags the sensor system, was used in [8]-[10]. Through tag motion, reported in this paper are significantly shorter than other line-to-resonator coupling is modulated, and consequently the previous tags with lower number of bits, used in similar near- transmission coefficient is modulated as well. Therefore, by field chipless-RFID systems [27]-[34]). feeding the line with a harmonic signal conveniently tuned (i.e., in the vicinity of the resonance frequency of the resonant II. WORKING PRINCIPLE AND DESIGN elements), the output signal is amplitude modulated (AM), and The principle of the proposed sensor and chipless-RFID the time distance between adjacent pulses, or dips, in the system is based on the variation of the transmission coefficient envelope function is related to the relative velocity between of a bandpass microstrip structure (particularly an order-1 the stator and the rotor (Fig. 1). Therefore, relative capacitively-coupled half-wavelength resonator bandpass instantaneous velocities as well as displacements between the filter) experienced when a straight metallic strip, etched in an stator and the rotor can be measured. Like in optical systems, a independent substrate, is located on top of it at short distance. figure of merit in these velocity and displacement sensors is Thus, in the proposed sensing/chipless-RFID system the active the number of pulses per time unit, intimately related to the part of the reader is a capacitively-coupled half-wavelength number of resonant elements per unit length. Thus, to increase resonator bandpass filter fed by a harmonic signal [see Fig. the number of pulses, it is necessary to reduce the period of 2(a)]. The transmission coefficient of the bare structure, i.e., the resonator chain as much as possible. without strip on top of it (but with the presence of the The main aim of this paper is to implement a linear substrate where the strips are etched), exhibits a resonance displacement and velocity sensor based on coupling peak at roughly the frequency where the length of the line is modulation, similar to those reported in [8]-[10] for angular equivalent to a half wavelength. Actually such frequency measurements, with a period significantly smaller. For this depends also on the pair of gaps present between the central purpose, a novel reader and chain are designed. To optimize line section and the input and output access lines, since such the chain period, very narrow (0.4 mm) and tiny separated (0.2 gaps provide a non-negligible reflection phase due to its mm) linear strips, oriented in the direction orthogonal to the capacitive effect. The resonance condition can be written as chain axis, are considered. By this means, we achieve a very 𝜑1 + 𝜑2 + 2𝛽𝑙1 = 2𝜋𝑛 (1) high density of strips per unit length, providing a high number where 1 and 2 are the reflection phases in gaps 1 and 2, of pulses per time unit (obviously, the time density of pulses is respectively, l1 is the electrical length of the central also determined by the relative velocity between the chain and transmission line section,  and l1 being the phase constant the reader). Moreover, as long as we can force the strip chain and the physical length of such section, and n = 1, 2, 3,.... to contain some inoperative strips, i.e., strips not detectable as From (1), it follows that for an uncoupled resonator (with 1 = pulses by the reader, we can use this system for coding 2 = 0), the first resonance (n = 1) occurs when l1 =  (or l1 = purposes. Within this scenario, each strip provides a bit of /2, where  is the guided wavelength), as it is well known. information with the logic state set to ‘1’ or ‘0’ depending on From (1), it is clear that the resonance condition can be whether the corresponding strip is operative or not. The strip altered either by modifying the reflection phases or by chain containing the code can be read through near-field by perturbing the phase constant (or both simultaneously). This means of the same reader used for measuring the linear can be achieved by the presence of a straight strip on top of displacement or velocity. In this case, however, only the the half-wavelength resonator, parallel oriented to the line pulses of those bits set to ‘1’ will appear in the envelope axis, with a length extending over the gap positions, as function. Since these chipless-RFID tags are implemented as indicated in Fig. 2(b). Figure 3 depicts the response of the bare chain patterns, reducing the chain period (as it is demonstrated 3 (a) RO4003C h = 8 mils εr = 3.55 (b) -5 Pin Pout Pin Pout -10 RO4003C h = 32 mils εr = 3.55 |S21| (dB)|f = f0 -15 (c) (d) Pin Pout Pin Pout -20 -25 Fig. 4. Cross-sectional view of the strips and reader corresponding to the four -30 Alternating operative and inoperative strips responses depicted in Fig. 3. (a) Bare reader; (b) loaded reader with operative All strips operative strip; (c) loaded reader with inoperative strip with a cut in the central position; -35 (d) loaded reader with shortened inoperative strip. 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Reader Circulator Envelope Displacement (mm) detector Δx/vr/Tag ID Fig. 6. Transmission coefficient at f0 as a function of the relative displacement Z0 … V0 (t) V0 ... between the reader and the strip chain. Two cases are considered, i.e., with all t0 Vi @ f0 R C strips operative, and with operative and inoperative (shortened) strips Z0 t0 tN-1 alternating. … tN-1 Encoder identical to the one of the resonant element of the reader. By this means, the effect of the strip on the gap capacitance is Fig. 5. Sketch illustrating the operation principle of the displacement/velocity expected to be smaller, and, consequently, a response closer to sensor and chipless-RFID system. the one of the bare reader is expected. This is confirmed in Fig. 3, where such response is also depicted. According to reader inferred from full wave electromagnetic simulation these results, tag programming should be done through the (using Keysight Momentum), as well as the response of the second considered approach, i.e., by shortening the strip, since reader with a perfectly aligned strip on top of it, where it can the resulting response, very similar to the one of the bare be appreciated that the resonance peak is shifted to lower reader, exhibits a significant excursion of the transmission frequencies. Actually, the responses of the bare and loaded coefficient at the resonance frequency of the reader loaded reader have been obtained by considering the presence of with operative strip, f0, indicated in Fig. 3. adjacent strips separated 0.2 mm (corresponding to a period of For the measurement of displacements and velocities, as 0.6 mm). The reason is that for the measurement of well as for tag reading in a near-field chipless-RFID scenario, displacements and velocities, a strip chain is necessary, as will the idea is to displace the strip chain above the reader at short be later shown. distance (the considered air gap in the results depicted in Fig. For the application of this structure (reader and strip chain) 3 is 0.2 mm). For proper tag reading or displacement/velocity to near-field chipless-RFID, it is necessary to make measurement, it is necessary that the strip chain is oriented inoperative those strips of the tag corresponding to the logic with the strips parallel to the reader axis. Moreover, the central state ‘0’. The most canonical option is to eliminate such strips position of the strips must coincide to a good approximation (so that bit encoding is determined by the presence or absence with the central position of the half-wavelength resonator of of strip). However, this is not the best option, in terms of cost, the reader. By tuning the feeding harmonic signal of the reader for massive tag manufacturing, since each code needs a to f0 (the resonance peak of the loaded reader), the amplitude different mask. Therefore, the preferred strategy in a real of the output signal is expected to be modulated by strip (or scenario is to fabricate all-identical tags (using massive tag) motion. From the envelope function, which can be printing processes such as screen-printing, rotogravure, inferred from an envelope detector preceded by an isolator (to flexography, etc.), and then personalize (program) the tags by avoid unwanted reflections from the diode), the relative means of low-cost systems, such as inkjet or laser ablation. velocity/displacement between the strip chain in motion and In [31]-[34], tag programming was achieved by cutting the fixed reader, or the identification (ID) code of the tag, can certain resonant elements in the tag chain (making them be inferred. The operation principle is illustrated in Fig. 5. inoperative). In the system presented in this work, we have For displacement and velocity measurements, all the strips considered two options to “write” the logic ‘0’ (strip should be present and functional for resolution optimization. inoperative). The first one is to cut the strip in the central Since the chain period is known (0.6 mm in our case), from position [see Fig. 4(c)]. The transmission coefficient, also the distance between adjacent peaks, or dips, in the envelope depicted in Fig. 3, is different to the one corresponding to the function, the velocity of the strip chain in motion can be presence of a functional strip on top of the reader, but it is far obtained. The relative displacement can be simply derived from the response of the bare reader. The reason is thought to from the number of cumulative peaks, or dips, appearing from be due to the fact that the main cause of resonance variation is a reference position. According to Fig. 3, the excursion the modification of the effective capacitance of the gap, which experienced by the transmission coefficient at f0 when the strip is not altered by implementing a cut in the central position of is made inoperative (by shortening it) is significant. However, the strip. The second approach consists of shortening the strips this does not ensure an important variation of the transmission by the extremes [see Fig. 4(d)], providing a strip length coefficient at that frequency by strip motion, if the strips are 4 tiny separated (as pursued in this work for resolution (a) -5 optimization). To analyze this important aspect, we have -10 obtained the transmission coefficient at f0 for different relative |S21| (dB)|f = f0 -15 displacements between the reader and the strip chain (where 5 operative strips separated 0.2 mm, providing a chain period of -20 0.6 mm, have been considered). -25 0 mm 0.6 mm The result is depicted in Fig. 6. It can be seen that the variation -30 0.2 mm 0.8 mm of the transmission coefficient is roughly 3.7 dB, which may -35 0.4 mm 1 mm be enough to achieve a non-negligible modulation index in the output signal of the reader line, and hence an envelope 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 function with readable (distinguishable) peaks, or dips. Displacement (mm) By alternating the functionality of the strips (i.e., the central and the outer strips operative and the other two shortened), (b) -5 roughly equivalent to duplicate the chain period, the excursion -10 |S21| (dB)|f = f0 experienced by the transmission coefficient is as high as 16.4 -15 dB (see Fig. 6). This later situation is much better in terms of -20 robustness, since a larger dynamic range in the envelope function is expected, but with the penalty of a degraded -25 resolution. For this main reason, we have maintained the -30 0 2 4 period of the strip chain in 0.6 mm in the fabricated chains (or 1 3 -35 tags). Obviously, intermediate periods can be considered, but, 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 as mentioned, the main aim of this paper is to drive the resolution to the limits imposed by the available technology Displacement (mm) (the period in the resonant chains of the sensors reported in Fig. 7. Effects of lateral misalignment (a) and lack of alignment by rotation [9],[10], based on a similar principle, is 2.2 mm). (b) on the transmission coefficient at f0 when the strip chain is displaced over the reader. The considered air gap is 0.2 mm. This small resolution for displacement and velocity (a) -5 measurements has direct impact on data density when the structure is used as chipless-RFID system. In this later case, -10 |S21| (dB)|f = f0 the relative velocity between the reader and the tag (the strip -15 chain) must be well known and stable during a reading -20 operation. From a mechanical viewpoint, implementing a guiding system providing a uniform tag speed in a real -25 0.2 mm 0.22 mm scenario is not an issue (nevertheless, this mechanical aspect is -30 0.21 mm 0.23 mm out of the scope of this paper, purely focused on the -35 electromagnetic and electronic blocks of the system). The 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 need to know the tag velocity with regard to the reader comes from the fact that such velocity determines the time instants Displacement (mm) where the envelope function must be read. In this regard, it (b) -5 should be mentioned that header bits with all strips operative -10 (‘1’) may be added at the beginning of the ID code in order to |S21| (dB)|f = f0 provide the information relative to the tag velocity. -15 Another important aspect related to robustness, is the -20 readability of the system when it is subjected to mechanical -25 vibrations or misalignments. Thus, we have studied the effects 0.2 mm 0.18 mm 0.16 mm -30 of misalignment/orientation and vertical distance (gap) 0.19 mm 0.17 mm variation between the reader and the strip chain on the -35 response of the system, particularly, the transmission 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 coefficient at f0. For that purpose, a case example has been Displacement (mm) considered, i.e., a 5-strip chain with alternating operative and Fig. 8. Effects of air gap on the transmission coefficient at f0 when the strip inoperative strips (identical to the one considered in Fig. 6). chain is displaced over the reader. (a) Air gaps above 0.2 mm; (b) air gaps Figure 7(a) depicts the simulated responses obtained by below 0.2 mm. displacing the strip chain over the reader for different lateral misalignments. It can be seen that the system is tolerant up to strips (with the penalty of larger size), the above tolerance in lateral displacements of roughly 0.8 mm (above that value, lateral displacement can be improved. the operative and inoperative strips cannot be distinguished). Let us now consider the effects of misalignment in relative Nevertheless, it should be mentioned that by elongating the orientation (rotation) between the strip chain and the reader. 5 can conclude that system tolerance against vertical distance is at least in the range 0.16-0.23 mm. It is worth mentioning that for air gaps of 0.16 mm and 0.17 gaps mm, a double peak in the response is visible. This is due to the fact that by decreasing the air gap, the maximum in the frequency response of the structure (coincident with f0 in Fig. 3, where the air gap is 0.2 mm) is displaced to the left. Under these conditions, two separated maxima are expected when the strip chain is displaced above the reader and the frequency of the injected signal is f0. Obviously, such maxima are visible (double peak) when the frequency of the maximum transmission coefficient lies significantly below f0, and this occurs for sufficiently small air gaps (such as 0.16 mm and (a) (b) 0.17 mm in our case). Fig. 9. Photograph of the strip chain (a) and active part of the reader (b). Vi @ f0 III. EXPERIMENTAL RESULTS AND DISCUSSION a. Linear displacement and velocity sensor We have fabricated a strip chain consisting of 100 all- Guide system identical operative strips by means of photo-etching on the substrate indicated in the caption of Fig. 2. The active part of Chipless RFID the reader (i.e., the capacitively coupled half-wavelength Tag resonator bandpass filter) has been fabricated by means of a milling machine LPKF-H100 on the substrate also indicated in the caption of Fig. 2. The photographs of the chain and reader are depicted in Fig. 9. The frequency of the feeding (carrier) signal has been set to f0 = 4 GHz, and such signal has been generated by means of the function generator Agilent E4438C. The envelope function V0 of the AM modulated signal present at the output port of the active part of the reader has been inferred by means of an Schottky envelope detector implemented by means of the Avago HSMS- Reader Isolator diode 2860 Schottky diode and the N2795A active probe (with capacitance 1 pF and resistance 1 M). Such envelope Fig. 10. Photograph of the experimental setup. detector has been preceded by an isolator (implemented by means of a circulator, model ATM ATc4-8), in order to avoid For that purpose, we have obtained the response inferred by mismatching reflections from the diode, a highly nonlinear displacing the strip chain over the reader for different angles device. The displacement of the strip chain over the active part between the line axis (reader) and the axis of the strips in the of the reader has been achieved by means of an in-house chain (the perfect situation corresponding to an angle of 0º). guiding system to which the chain has been attached. The The results, depicted in Fig. 7(b), indicate that the system is photograph of the experimental setup is shown in Fig. 10. tolerant up to an angle of roughly 3º, which is reasonable After adequately orienting and positioning the strip chain from a practical viewpoint. Better tolerance against lack of over the reference position, we have displaced the chain, and alignment by rotation can be obtained, but at the expense of we have recorded the envelope function in an oscilloscope wider line and strips (which represents a penalty in terms of (model Agilent MSO-X-3104A). The resulting envelope data density for chipless-RFID systems and resolution for function is depicted in Fig. 11. Note that this function is not linear displacement measurements). exactly periodic. This is mainly due to the non-uniformity of The effects of the air gap on the response of the system can the air gap during the displacement of the strip (note also that be seen in Fig. 8. The structure has been optimized for a our guiding system cannot guarantee a uniform velocity, as nominal air gap of 0.2 mm, where the maximum excursion of well). Nevertheless, from the distance between adjacent peaks, the transmission coefficient results. Increasing the air gap or dips, the instantaneous linear velocity can be obtained. [Fig. 8(a)] reduces such excursion, but reasonable values are According to the figure, the dips (appearing when the obtained within the considered margin (up to 0.23 mm). intermediate position between two strips is perfectly aligned Figure 8(b) shows the responses that have been inferred by with the line axis of the reader) are clearly visible and easily decreasing the air gap. It is shown that even for a vertical identifiable. Such dip times, which can be easily recorded by distance between the strip chain and the reader as small as means of a post-processing unit, are then used to determine the 0.16 mm, the excursion experienced by the transmission coefficient is reasonable. Hence, according to these results, we 6 1.0 2.0 1.0 1 0 1 0 1 0 V0/Vmax 0.9 1.5 0.5 V0/Vmax 1.0 0.0 0.8 0.5 0.7 Tm  0.055 s vm  1.091 cm/s 0.0 0.6 4.0 4.1 4.2 4.3 4.4 2.0 1.0 1 1 1 0 1 1 V0/Vmax Time (s) 1.5 0.5 Fig. 11. Measured normalized envelope function obtained by displacing the V0/Vmax strip chain over the reader (only a portion of the curve is shown for a better view). 1.0 0.0 linear velocity and the relative displacement between the 0.5 reader and the chain from the reference position. According to Fig. 11, the average time interval between adjacent dips is 0.0 0.055 s, corresponding to an average displacement velocity of 2.0 1.0 1 0 1 0 0 0 V0/Vmax 1.09 cm/s. 1.5 0.5 The main advantage of this system for the measurement of linear displacements and velocities is the fact that it exhibits a 0.0 1.0 large input dynamic range and, simultaneously, very good resolution. Particularly, the resolution for linear displacement 0.5 measurement is dictated by the period, which is only 0.6 mm. This is a very small resolution if we take into account that the 0.0 input dynamic range is only limited by the total length of the 13 14 15 16 17 strip chain. Therefore, this system may be of interest for Time (s) motion control, in applications where sub-millimeter Fig. 12. Measured normalized envelope function corresponding to the tags resolution and dynamic ranges of several centimeters or even with the indicated codes (shown in the insets, where zoom views are included meters are needed. In displacement sensors based on a for clarity). transmission line loaded with a single resonant element (e.g., those reported in [2]-[6]) the resolution can be driven to lower (although the peaks and dips can be perfectly discerned), values, but the dynamic range is roughly limited by the size of which is due to the fact that the resolution has been driven to the resonant element, i.e., far from the dynamic ranges the limits, as discussed before. Note that by increasing achievable with the approach reported in this work. somehow the period of the strip chain, yet preserving a good For the measurement of instantaneous velocities, the key resolution, the ratio between the dip and peak levels in the figure of merit is the number of pulses per time unit, dictated envelope function is expected to be significantly improved by the period of the strip chain (and obviously by the motion (reduced). Note, for instance, that with a period of 1.2 mm speed). In this work, we have achieved a substantially smaller (equivalent to alternating operative and inoperative strips in a period than those of other previously reported displacement 0.6mm-period chain), the variation in the transmission and velocity sensors [9],[10] (applied to the measurement of coefficient is as large as 16.4 dB (see Fig. 6). Consequently, a angular velocities). This has been achieved by simply smaller ratio between the dip and peak levels in the envelope replacing the rectangular resonant elements of the chains used function should be expected in this case. in [9],[10] by straight strips (orthogonally oriented to the chain Concerning the maximum velocity that can be measured axis), and, obviously, by designing a new reader, able to detect with the proposed system, in a real scenario it is dictated by the functionality, or not, of such strips. By this means, the the sampling frequency of the acquisition system, fs. For period has been reduced from 2.2 mm in [9],[10] to 0.6 mm in reasonable values of fs, extremely high velocities can be the sensor reported in this work. measured. For instance, if we consider a minimum number of Note that the variation experienced by the envelope function samples per cycle of the envelope function (distance between in Fig. 11 reasonably correlates with the variation of the adjacent pulses or dips) and this number is set to 100 samples transmission coefficient in Fig. 6. According to this figure, the per cycle, then it follows that the time period (distance difference between the minima and the maxima of the between adjacent pulses, tp) should be not less than transmission coefficient at f0 (when all the strips are operative) 100/fs, i.e., tp > 100/fs. Therefore, the velocity must satisfy vp is roughly −3.7 dB, which corresponds to a power ratio of 0.42 < p/tp, where p is the period of the encoder (p = 0.6 mm in our and to a voltage ratio of 0.65, close to the ratio between the case), and the maximum velocity that can be measured, by dip and peak levels in Fig. 11. This is a relatively high ratio considering a reasonable sampling frequency of fs = 100 MHz, is found to be 600 m/s, which is more than enough for most 7 applications. Concerning the lower limit, velocities as small as fabricated, and then programmed by elongating the required can be measured, obviously, assuming that at least metallization of certain strips (those which should be set to two adjacent pulses (or dips) are needed to determine the ‘1’) by means of a printing process (e.g., inkjet printing). The relative velocity between the tag and the reader. normalized envelope functions and the generated codes are With regard to velocity resolution, it is related to time depicted in Fig. 12. It can be seen that the tags are correctly resolution, given by the inverse of the sampling frequency read by virtue of the maxima and minima associated to the (i.e., tres = 1/fs). Since the velocity can be expressed as logic ‘1’ and ‘0’ states, respectively. Note that in the tag with 𝑝 𝑝 𝑡𝑟𝑒𝑠 sequence ‘101010…’ the average ratio between the minimum 𝑣𝑝 = ≈ (1 ∓ ), (2) and the maximum in the envelope function is very small, in 𝑇𝑚±𝑡𝑟𝑒𝑠 𝑇𝑚 𝑇𝑚 agreement with the 16.4 dB variation in the transmission Tm being the time distance between adjacent pulses, it follows coefficient at f0 inferred by displacing the tag (see Fig. 6). that the velocity resolution is given by Such envelope function confirms the smaller achievable ratios 𝑝𝑡 𝑡𝑟𝑒𝑠 𝑡 𝑣𝑝,𝑟𝑒𝑠 = 𝑟𝑒𝑠2 = 𝑣𝑝 = 𝑣𝑝2 𝑟𝑒𝑠 (3) between dips and peaks when the “effective” chain period is 𝑇𝑚 𝑇𝑚 𝑝 increased, as anticipated before. and it depends on the linear velocity. To give an idea of Conceptually, the present near-field chipless-RFID system is specific values, the velocity resolution resulting from (3), by similar to those reported in previous papers [27]-[34]. considering a nominal velocity of 0.1 m/s (with p = 0.6 mm However, the main benefit of the proposed system is the fact and fs = 100 MHz), is as small as 1.6610-7 m/s. By that the tag period has been substantially reduced. This means considering a higher velocity, e.g., 100 m/s, the resolution is that the per-unit length data density, as high as 16.66 bits/cm, found to be 0.166 m/s. Note that for a given period and is substantially larger than in the tags reported, e.g., in [31]- sampling frequency, or tres, the resolution varies quadratically [33] (specifically, more than 5 times larger). Indeed, the with the nominal velocity. This explains the degradation of the proposed 100-bit tag is as short as 6 cm. By contrast, the 84- resolution with velocity. Indeed, a relevant parameter is the bit tag reported in [33] exhibits a length of 26.7 cm ratio between the resolution and velocity, i.e., (corresponding to a per unit length data density of 3.14 𝑣𝑝,𝑟𝑒𝑠 𝑡𝑟𝑒𝑠 bit/cm). Obviously, by using straight strips orthogonally = 𝑣𝑝 (4) 𝑣𝑝 𝑝 oriented with regard to the chain axis, the width of the which is found to be proportional to the nominal velocity. proposed tags (2.24 cm) is larger than the one of the tags Another important parameter related to sensor performance reported in [33] (0.35 cm). However, in these tags based on is the error in the determination of the linear velocity. As long near-field coupling and sequential bit reading, implemented as as the time period in this work is determined from the distance chains of metallic patterns, the limiting factor (when a large between adjacent dips (see Fig. 11), such error depends on dip number of bits is necessary) is dictated by tag length, provided width and distance between dips. As it was pointed out in the width is constrained within reasonable values. Therefore, [9],[10], narrower dips provide smaller error. Nevertheless, the the proposed approach represents a good solution in this accuracy can be improved either by averaging over regard. The data capacity of the presented tags (100 bits) is by several dips, or by considering distant dips (assuming in far larger than the data capacity reported to date in other both cases a degradation in the ability of the sensor to chipless-RFID tags (an exhaustive comparative analysis of the detect velocity variations in small time lapses). Dip width different chipless-RFID technologies can be found in [34]). can be estimated by a predefined percentage variation from the Another important difference between the present system minima, e.g., 2%. With this (conservative) value, the ratio and those reported in [27]-[34] concerns the fact that in the between the dip width and distance between adjacent dips is structures reported in the above references, the tag chain is found to be 0.03, corresponding to a time error of 3% and composed of resonant elements, and such elements are roughly to the same velocity error. Note, however, that such identical to the one present in the reader (either S-shaped split error can be reduced by an order of magnitude by simply ring resonators, S-SRRs, or simply SRRs). In this paper, the determining the velocity from the time interval corresponding elements of the tag (straight strips) modify the resonance to 10 dips. frequency of the reader line (an order-1 bandpass filter), thereby modulating the amplitude of the interrogation signal, b. Chipless-RFID system but such elements do not act as resonant elements. Note also that the microwave part of the reader, the one needed to detect To demonstrate the potential of this system for near field- the presence of operative or inoperative strips through the chipless-RFID applications, we have programmed three codes amplitude of the envelope function, is completely different as from the single fabricated 100-bit tag with all bits set to ‘1’, compared to papers [27]-[34]. i.e., the strip chain shown in Fig. 9(a). As mentioned before, The main focus of this paper has been to decrease the period the different codes have been generated by slightly shortening of the tag chain as much as possible in order to improve the those strips associated to the logic state ‘0’. We have done this resolution in the displacement sensor, and the density of bits by a milling procedure, but processes such as laser ablation per unit length in the chipless-RFID system. We have can be envisaged in a real scenario. Alternatively, all-identical achieved a period of 0.6 mm, much smaller than those tags with all bits set to ‘0’ (i.e., with shortened strips) can be reported in the chipless-RFID systems of [27]-[34] and in the 8 angular velocity sensors of [8]-[10]). Such small period has the proposed system for chipless-RFID applications is the been achieved thanks to the new combination of encoder and achievable data capacity (100-bit tags have been reported), reader topology, able to detect the dips/pulses with such small compatible with small-sized tags, as short as 6 cm in the separation between the tag elements. Moreover, although the present work. This high per unit length data capacity, the most excursion experienced by the transmission coefficient (as important figure of merit in chipless-RFID tags based on near- consequence of a bit variation in the ID coding sequence) is field coupling and sequential bit reading, has been achieved, larger in [31]-[33], it can be appreciated in Fig. 12 that the again, thanks to the small period of the designed and different codes can be perfectly read. fabricated tags. The reported devices may be of interest for It is worth mentioning that a properly encoded tag (e.g., with accurate control motion, and for chipless-RFID, in a non-symmetric periodic sequence) may be used as linear applications where the read distance can be sacrificed in favor displacements and velocity sensor, with the capability to of a high data capacity (e.g. secure paper). To the best of our detect the direction of motion. Note, however, that motion knowledge, 100-bit chipless-RFID tags have been never direction detection with this procedure has the penalty of a reported to date. worst displacement resolution, due to the reduction in the number of pulses. Also, the capability to measure REFERENCES instantaneous velocities is degraded, since variations in time [1] M. Schueler, C. Mandel, M. Puentes, and R. Jakoby, “Metamaterial lapses smaller than the encoder period cannot be detected. inspired microwave sensors,” IEEE Microw. Mag., vol. 13, no. 2, pp. 57–68, Mar. 2012. Nevertheless, in its functionality as linear displacement and [2] C. Mandel, B. Kubina, M. Schüßler, and R. Jakoby, “Passive chipless velocity sensor, either with all functional resonators or with an wireless sensor for two-dimensional displacement measurement,” Proc. encoded sequence, the proposed structure may be useful for 41st European Microwave Conf., Manchester, UK, 2011, pp. 79–82. many motion control applications, where the velocity does not [3] J. Naqui, M. Durán-Sindreu and F. Martín, “Novel Sensors Based on the Symmetry Properties of Split Ring Resonators (SRRs)”, Sensors, vol 11, experience strong variations with time. pp. 7545-7553, 2011. [4] A. Karami-Horestani, C. Fumeaux, S.F. Al-Sarawi, and D. Abbott, IV. CONCLUSIONS “Displacement sensor based on diamond-shaped tapered split ring resonator,” IEEE Sensors J., vol. 13, no. 4, pp. 1153-1160, Apr. 2013. In conclusion, a novel near-field approach useful for the [5] A. Horestani, D. Abbott, and C. Fumeaux, “Rotation sensor based on measurement of linear displacements and instantaneous horn-shaped split ring resonator,” IEEE Sens. J., vol. 13, no. 8, pp. 3014–3015, May 2013. velocities, as well as for the implementation of chipless-RFID [6] A. Ebrahimi, W. Withayachumnankul, S. Al-Sarawi, and D. Abbott, systems has been presented in this paper. The approach is “Metamaterial-inspired rotation sensor with wide dynamic range,” IEEE based on a periodic chain of straight metallic strips etched on a Sensors J., vol. 14, no. 8, pp. 2609-2614, Aug. 2014. [7] J. Naqui, and F. Martín, “Angular displacement and velocity sensors dielectric substrate. Such strips chain acts as the mobile based on electric-LC (ELC) loaded microstrip lines”, IEEE Sensors J., element in the displacement/velocity sensor, or as the element vol. 14(4), pp. 939-940, Apr. 2014. containing the ID code (or tag) in the chipless-RFID system. [8] J. Naqui, F. Martín, “Application of broadside-coupled split ring In this later case, however, certain strips should be slightly resonator (BC-SRR) loaded transmission lines to the design of rotary encoders for space applications“, IEEE MTT-S Int. Microw. Symp. shortened (making them inoperative), according to the (IMS’16), San Francisco (CA), May 2016. required bit code, since each strip in the chain corresponds to a [9] J. Mata-Contreras, C. Herrojo, and F. Martín, “Application of split ring single bit element. For tag reading, an element able to resonator (SRR) loaded transmission lines to the design of angular displacement and velocity sensors for space applications”, IEEE Trans. distinguish the longer and shorter strips is required, and for Microw. Theory Techn., vol. 65, no. 11, pp. 4450-4460, Nov. 2017. that purpose, a capacitively coupled half-wavelength resonator [10] J. Mata-Contreras, C. Herrojo, and F. Martín, “Detecting the rotation order-1 bandpass filter (the active part of the reader), fed by a direction in contactless angular velocity sensors implemented with rotors loaded with multiple chains of split ring resonators (SRRs)”, IEEE harmonic signal conveniently tuned, has been used. Such Sensors J., vol.18, no. 17, pp. 7055-7065, Sep. 2018. element has been also used for displacement and velocity [11] M. S. Boybay and O. M. Ramahi, “Material characterization using measurements, since, with additional electronics, it is able to complementary split-ring resonators,” IEEE Trans. Instrum. Measur., vol. 61, no. 11, pp. 3039–3046, 2012. provide the time interval between two adjacent strips crossing [12] C.-S. Lee and C.-L. Yang, “Complementary split-ring resonators for a fixed point (the reader), and hence the relative velocity and measuring dielectric constants and loss tangents,” IEEE Microw. displacement. For such measurements, and for tag reading, the Wireless Compon. Lett., vol. 24, no. 8, pp. 563−565, Aug. 2014. strip chain is displaced above the active part of the reader, in [13] C.-L. Yang, C.-S. Lee, K.-W. Chen, and K.-Z. Chen, “Noncontact measurement of complex permittivity and thickness by using planar close proximity to it. By this means, the output signal is resonators,” IEEE Trans. Microw. Theory Techn., vol. 64, no.1, pp. amplitude modulated due to the effects of the operative strips 247−257, Jan. 2016. on the transmission coefficient at the frequency of the feeding [14] W. Withayachumnankul, K. Jaruwongrungsee, A. Tuantranont, C. Fumeaux, and D. Abbott, “Metamaterial-based microfluidic sensor for signal, and the information relative to the displacement and dielectric characterization,” Sensor Actuat. A Phys., vol. 189, pp. 233- velocity, as well as the ID code, is contained in the envelope 237, Jan. 2013. function. As sensing element, the reported device exhibits sub- [15] A. Ebrahimi, W. Withayachumnankul, S. Al-Sarawi, and D. Abbott, “High-sensitivity metamaterial-inspired sensor for microfluidic dielectric millimeter resolution (0.6 mm) and high input dynamic range characterization,” IEEE Sensors J., vol. 14, no. 5, pp. 1345-1351, May for linear displacement measurement. Moreover, by virtue of 2014. the small chain period, the number of pulses per unit time, a [16] A. Abduljabar, D. Rowe, A. Porch, and D. Barrow, “Novel microwave figure of merit in sensors devoted to the measurement of microfluidic sensor using a microstrip split-ring resonator,” IEEE Trans. Microw. Theory Tech., vol. 62, no. 3, pp. 679-688, Mar. 2014. instantaneous velocities, is very high. The main advantage of 9 [17] T. Chretiennot, D. Dubuc, and K. Grenier, “A microwave and Cristian Herrojo was born in Barcelona, Spain, in microfluidic planar resonator for efficient and accurate complex 1983. He received the Telecommunications Technical permittivity characterization of aqueous solutions,” IEEE Trans. Engineering degree in electronic systems and Microw. Theory Techn., vol. 61, no. 2, pp. 972–978, Feb. 2013. Telecommunications Engineering degree from the [18] A. Ebrahimi, W. Withayachumnankul, S. F. Al-Sarawi and D. Abbott, Universitat Autònoma de Barcelona in 2010 and 2012, "Microwave microfluidic sensor for determination of glucose respectively and the PhD degree in Electronics concentration in water," 2015 IEEE 15th Mediterranean Microwave Engineering from the same university in 2018. His Symposium (MMS), Lecce, 2015, pp. 1-3. research interests include RF/microwave devices, [19] A. Salim and S. Lim, “Complementary split-ring resonator-loaded Chipless-RFID and RFID technology, and microfluidic ethanol chemical sensor,” Sensors, vol. 16, pp. 1-13, 2016. Metamaterials. [20] P. Vélez, L. Su, K. Grenier, J. Mata-Contreras, D. Dubuc, and F. Martín, He has authored or co-authored over 12 journal papers and 10 conference “Microwave microfluidic sensor based on a microstrip splitter/combiner papers. He has also filed one patent on Chipless-RFID. configuration and split ring resonators (SRR) for dielectric characterization of liquids”, IEEE Sensors J., vol. 17, pp. 6589-6598, Oct. 2017 Francisco Javier Muela, was born in Formentera, [21] M. H. Zarifi, M. Fayaz, J. Goldthorp, M. Abdolrazzaghi, Z. Hashisho, Spain in 1988. He is studying Telecommunications and M. Daneshmand, “Microbead-assisted high resolution Technical Engineering degree in Electronic Systems microwave planar ring resonator for organic-vapor sensing,” Appl. at Universitat Autonoma de Barcelona, and he will Phys. Lett., vol. 106, no. 6, p. 62903, Feb. 2015. present the Bachelor Thesis on September 2018. [22] M. Abdolrazzaghi and M. Daneshmand, “Compelling impact of intermodulation products of regenerative active resonators on sensitivity,” in IEEE MTT-S Int. Microw. Symp. Dig., Jun. 2017, Honolulu (HI), pp. 1018–1021. [23] M. Abdolrazzaghi and M. Daneshmand, “Dual active resonator for dispersion coefficient measurement of asphaltene nano-particles,” Javier Mata-Contreras was born in 1976 in Málaga IEEE Sensors J., vol. 17, no. 22, pp. 7248–7256, Nov. 2017. (Spain). He received the Ingeniería de [24] M. H. Zarifi et al., “A microwave ring resonator sensor for early Telecomunicación Degree from the Universidad de detection of breaches in pipeline coatings,” IEEE Trans. Ind. Electron., Málaga (UMA) in 2000 and the PhD degree from the vol. 65, no. 2, pp. 1626–1635, Feb. 2018. same university in 2010, with the Thesis “Distributed [25] M. Abdolrazzaghi, M. H. Zarifi. W. Pedrycz, and M. Daneshmand, Amplifiers and Mixers with Transmission Lines based “Robust ultra-high resolution microwave planar sensor using fuzzy on Metamaterials”. In 2000, he joined the UMA neural network approach,” IEEE Sensors J., vol. 17, no. 2, pp. 323–332, Department of Ingeniería de Comunicaciones UMA as Jan. 2017. Assistant Professor. He is currently working at [26] F. Martín, F. Falcone, J. Bonache, R. Marqués and M. Sorolla, “Split CIMITEC and the Universitat Autònoma de Barcelona ring resonator based left handed coplanar waveguide”, Appl. Phys. Lett., as Visitant Professor. His research interests include active and passive vol. 83, pp. 4652-4654, Dec. 2003. microwave devices and active distributed circuits based on metamaterials, [27] C. Herrojo, J. Mata-Contreras, F. Paredes, Ferran Martín, “Near-field among others. chipless RFID encoders with sequential bit reading and high data capacity”, IEEE MTT-S Int. Microw. Symp. (IMS’17), Honolulu, Hawaii, June 2017. Ferran Paredes was born in Badalona (Barcelona), [28] C. Herrojo, J. Mata-Contreras, F. Paredes, F. Martín, “Microwave Spain in 1983. He received the Telecommunications encoders for chipless RFID and angular velocity sensors based on S- Engineering Diploma (specializing in Electronics) and shaped split ring resonators (S-SRRs)”, IEEE Sensors J., vol. 17, pp. the Telecommunications Engineering degree from the 4805-4813, Aug. 2017. Universitat Autònoma de Barcelona in 2004 and 2006, [29] C. Herrojo, J. Mata-Contreras, F. Paredes, A. Núñez, E. Ramón, F. respectively and the PhD degree in Electronics Martín, “Near-field chipless-RFID tags with sequential bit reading Engineering from the same university in 2012. He was implemented in plastic substrates”, Int. J. Magnetism. Magnetic Mat., Assistant Professor from 2006 to 2008 at the vol. 459, pp. 322-327, 2018. Universitat Autònoma de Barcelona, where he is [30] C. Herrojo, J. Mata-Contreras, F. Paredes, F. Martín, “High data density currently working as a Research Assistant. His and capacity in chipless radiofrequency identification (chipless-RFID) research interests include metamaterial concepts, tags based on double-chains of S-shaped split ring resonators (S-SRRs)”, passive microwaves devices, antennas and RFID. EPJ Appl. Metamat., vol. 4, article 8, 6 pages, Oct. 2017. [31] C. Herrojo, J. Mata-Contreras, F. Paredes, Ferran Martín, “Near-field Ferran Martín (M’04-SM’08-F’12) was born in chipless RFID system with high data capacity for security and Barakaldo, Spain, in 1965. He received the B.S. authentication applications”, IEEE Trans. Microw. Theory Techn., vol. degree in physics and Ph.D. degree from the 65 (12), pp. 5298-5308, Dec. 2017. Universitat Autònoma de Barcelona (UAB), [32] C. Herrojo, J. Mata-Contreras, F. Paredes, A. Núñez, E. Ramon, and F. Barcelona, Spain, in 1988 and 1992, respectively. Martín “Near-field chipless-RFID system with erasable/programmable From 1994 to 2006 he was Associate Professor in 40-bit tags inkjet printed on paper substrates”, IEEE Microw. Wireless Electronics at the Departament d’Enginyeria Compon. Lett., vol. 28, pp. 272- 274, Mar. 2018. Electrònica (Universitat Autònoma de Barcelona), and [33] C. Herrojo, J. Mata-Contreras, F. Paredes, A. Núñez, E. Ramon, and F. since 2007 he is Full Professor of Electronics. His Martín, “Very Low-Cost 80-Bit Chipless-RFID Tags Inkjet Printed on research activity has been very broad, including the Ordinary Paper,” Technologies, vol. 6, no. 2, p. 52, May 2018. modelling and simulation of electron devices for high frequency applications, [34] C. Herrojo, F. Paredes, J. Mata-Contreras, E. Ramon, A. Núñez, F. millimeter wave and THz generation systems, the application of Martín, “Time-domain signature barcodes: near-field chipless-RFID electromagnetic bandgaps to microwave and millimeter wave circuits, and the systems with high data capacity”, IEEE Microwave Magazine, to be application of metamaterial concepts to the miniaturization and optimization published. of microwave circuits and antennas. He is now very active in the development of microwave sensors for dielectric characterization and motion control, and also in the topic of chipless-RFID. He is the head of the Microwave Engineering, Metamaterials and Antennas Group (GEMMA Group) at UAB, and director of CIMITEC, a research Center on Metamaterials supported by TECNIO (Generalitat de Catalunya). He has organized several international events related to metamaterials, including Workshops at the IEEE International Microwave Symposium (years 2005 and 2007) and European 10 Microwave Conference (years 2009, 2015, 2017, and 2018), and the Fifth International Congress on Advanced Electromagnetic Materials in Microwaves and Optics (Metamaterials 2011), where he has acted as chair of the Local Organizing Committee. He has acted as Guest Editor in several Special Issues, mainly related to Metamaterials, in various International Journals. He has authored and co-authored over 580 technical conference, letter, journal papers and book chapters, he is co-author of the book on Metamaterials entitled Metamaterials with Negative Parameters: Theory, Design and Microwave Applications (John Wiley & Sons Inc. 2008), author of the book Artificial Transmission Lines for RF and Microwave Applications (John Wiley & Sons Inc. 2015), co-editor of the book Balanced Microwave filters (John Wiley & Sons Inc. and IEEE-Press 2018), and he has generated 19 PhDs. Ferran Martín has filed several patents on metamaterials and has headed several Development Contracts. Prof. Martín is a member of the IEEE Microwave Theory and Techniques Society (IEEE MTT-S). He is reviewer of the IEEE Transactions on Microwave Theory and Techniques and IEEE Microwave and Wireless Components Letters, among many other journals, and he serves as member of the Editorial Board of IET Microwaves, Antennas and Propagation and International Journal of RF and Microwave Computer-Aided Engineering. He is also a member of the Technical Committees of the European Microwave Conference (EuMC) and International Congress on Advanced Electromagnetic Materials in Microwaves and Optics (Metamaterials). Among his distinctions, Ferran Martín has received the 2006 Duran Farell Prize for Technological Research, he holds the Parc de Recerca UAB – Santander Technology Transfer Chair, and he has been the recipient of two ICREA ACADEMIA Awards (calls 2008 and 2013). He is Fellow of the IEEE since 2012 and Fellow of the IET since 2016.