Hussein Ezzeddine, Yanis Merakeb, Julien Huillery, Arnaud Bréard, Yvan Duroc
©SHUTTERSTOCK.COM/ALBERT LOZANO
Despite its first applications in the 1940s and 1950s, RF identification (RFID) technology only really developed starting in the early 2000s with the development of the international Electronic Product Code standard. As its name indicates, the original function of RFID was the identification of objects, animals, and people via a tag that contains a unique identifier. However, RFID is now being given new capabilities, such as integrated sensors or actuators, additional memory, and better security, which implies that its potential fields of application will keep growing, especially with the rise of the Internet of Things. In its passive version, which is the original RFID, the tag is a passive electronic component, mainly composed of an antenna and an integrated circuit (called an RFID chip), that is remotely powered and that emits its identifier using the principle of modulated backscattered communication. The remote powering and backscattering require a reader in the vicinity of the tag that acts as both a radio transmitter and a receiver and as an external RF power source.
The development of RFID is associated with the progress of microelectronics, and for a long time one of the main challenges for passive ultrahigh frequency (UHF) RFID chip manufacturers was to decrease the activation power of the chips (i.e., their sensitivity) to increase the reading distances. With the decrease in the sensitivity of the chips, a second issue appeared: that of the sensitivity of reception of the reader. Indeed, with the distance of activation of the tag increasing, the retro-modulated signal is all the more attenuated during the backward link propagation. It is therefore more difficult for the reader to distinguish the two states of information returned by the tag.
To the sensitivity of the tag (i.e., activation of the active circuits of the tag) and to the sensitivity of the reader (i.e., performance of the reception of the RF signals), which are related to the performance of the electronics, a third issue has been added: that of the impact of the medium in which the waves are propagating, reflecting the environmental context of the application. (As discussed later, this issue is perhaps even more important in the field of RFID than in general communication systems, as the signals encounter the channel twice.) Indeed, for applications that take place indoors and in dense environments in general (e.g., warehouses, buildings, and factories), communications are subject to small-scale multipath interferences in addition to path losses. Small-scale multipath effects cause rapid variations in signal power that are observed over a distance on the order of a wavelength. Multipath interference causes frequency-selective attenuation that is highly dependent on the propagation channel. In addition, the presence of a direct path [i.e., line-of-sight (LOS) propagation] is not guaranteed, and often the channel is more like a non-LOS (NLOS) path. As a result, a multipath channel affects not only the signal strength but also its amplitude and phase, which can greatly distort the signal [1], [2].
In the context of passive RFID, modulated backscattered signal attenuation due to small-scale multipath effects is the product of attenuation of both the forward link and the backward (backscatter) link [3], [4]. In the case of the forward link, multipath interference can reduce the power received by the tag below the threshold it requires to activate and/or increase the read cycle time of a tag or a population of tags. In real scenarios of complex environments, part of the tags, although located at theoretically reachable distances, may not be read by a commercial reader [5], [6]. As a result, the effects of multipath interference degrade the overall performance of RFID systems, that is, the range and reliability of communications.
While some passive RFID tags are designed to target specific types of applications (e.g., miniature tags or tags dedicated to metal objects), in complex propagation environments the deployment of an RFID solution can quickly become limited. In many situations the complexity of the channel can become an important obstacle, leading to impossible or highly degraded tag readings. Thus, the need for a certain robustness with respect to propagation channels (and thus indirectly to applications) will become increasingly important for passive RFID technology. In recent years, it has been demonstrated that an alternative to the use of the “conventional” carrier wave (i.e., a sine wave at the operating frequency) is to consider specific RF waveforms that are optimized for the application under consideration [7], [8]. Indeed, considering the properties of the propagation channel and/or the characteristics of the tag’s electronic circuits in the design of RF waveforms makes it possible to optimize both energy and information transfer performance. In the field of RFID, there are some studies, still quite rare, which attempt to improve the performance of passive RFID communications by exploiting optimal waveforms [9], [10]. This work aims to optimize the forward link from the reader to the tag by improving the wireless power transfer. For a global optimization of an RFID system, it is also imperative to take into account the backward link from an energetic point of view as well as by considering the information signal backscattered by the tag. In a complex propagation environment, this low-level signal can be difficult to detect at the reader. In light of these two challenges, this article summarizes the work we have developed at the Ampère research laboratory in France on the potential of passive UHF RFID in pulsed wave (PW) mode.
In the classical continuous wave (CW) mode, the signal transmitted by the reader is a monochromatic wave given by \[{x}_{\text{CW}}{\left({t}\right)} = {A}\,{\cos}\,{\left({2}{\pi}{f}_{c}{t}\right)} \tag{1} \] where ${A}$ is the amplitude and ${f}_{c}$ is the carrier frequency. The principle behind PW mode RFID is to replace this traditional monochromatic signal with a regular train of pulses with a given period ${T}_{0}$. Considering a pulse ${p}{\left({t}\right)}$ with a finite duration and a bandwidth ${B}$ centered around a carrier frequency ${f}_{c}$, the PW mode signal emitted by the reader can be written as \[{x}_{\text{PW}}{\left({t}\right)} = {p}{(}{t}{)}\,{\ast}\,{\text{III}}_{T_0}{(}{t}{)} \tag{2} \] where IIIT0( . ) denotes the Dirac comb with parameter ${T}_{0}$, and ${\ast}$ is the convolution product. Because the pulse ${p(t)}$ has a finite bandwidth ${B}$, its spectrum ${P}{\left({f}\right)}$ is null for all frequencies outside the interval ${\left[{f}_{c}{-}{\left({B} / {2}\right)}{;}\,{f}_{c} + {\left({B} / {2}\right)}\right]}$ and, considering only the positive frequencies, the Fourier transform of ${X}_{\text{PW}}{\left({f}\right)}$ can be written as \[{X}_{\text{PW}}{\left({f}\right)} = \frac{1}{{T}_{0}} \mathop{\sum}\limits_{{n} = {n}_{0}}\limits^{{n}_{0} + {N}{-}{1}}{P}{\left(\frac{n}{{T}_{0}}\right)}{\delta}{\left({f}{-} \frac{n}{{T}_{0}}\right)} \tag{3} \] where ${\delta}{\left({\cdot}\right)}$ denotes the Dirac impulse. From (3), it is clear that the signal ${x}_{\text{PW}}{\left({t}\right)}$ is a sum of ${N}$ harmonic components and can therefore also be written as \[{x}_{\text{PW}}{\left({t}\right)} = {\frak{R}}{e}{\left[\mathop{\sum}\limits_{{n} = {0}}\limits^{{N}{-}{1}}{\alpha}_{n}{e}^{{i}{2}{\pi}{f}_{n}{t}}\right]} \tag{4} \] where ${\frak{R}}{e}{\left[{.}\right]}$ denotes the real part, and the frequencies ${f}_{n}$ and the complex coefficient ${\alpha}_{n}$ are defined as ${\forall}{n}\,{\in}\,{\left\{{0},{\ldots},{N}{-}{1}\right\}}$, ${f}_{n} = {\left({n}_{0} + {n} / {T}_{0}\right)}$, and ${\alpha}_{n} = {\left({1} / {T}_{0}\right)}{P}{\left({f}_{n}\right)}$.
The two formulations given by (2) and (4) are strictly equivalent. The former highlights the impulsive nature of the PW mode signal; the latter brings more insight into its multifrequency content (as opposed to the CW mode), which must be taken into account to optimally recover the information backscattered by the tag.
The RF signal ${r(t)}$ received at the reader antenna after backpropagation from the tag can be expressed as \[{r}{\left({t}\right)} = {\frak{R}}{e}{\left[{C}\,{cdot}\,{(}{1} + {\gamma}\,{\cdot}\,{m}{(}{t}{)}\right]} \tag{5} \] where ${x(t)}$ is the signal transmitted by the reader (regardless of whether it is in CW or in PW mode), ${m}{(}{t}{)}\,{\in}\,{\left\{{-}{1},{1}\right\}}$ denotes the binary signal representing the FM0-coded binary sequence backscattered by the tag, ${C}$ is a complex gain representing the attenuation and the phase due to the propagation, and ${\gamma}$ is a complex gain representing the difference between the two reflection coefficients of the tag in the low and high states (note that ${\gamma}$ is proportional to the differential radar cross section introduced in [11]). The objective of the processing carried out by the reader is to reconstruct ${m(t)}$ from the RF signal ${r(t)}$. In the case of the CW mode where ${x(t)}$ is given by (1), the signal received takes the following form: \[{r}_{\text{CW}}{\left({t}\right)} = {\frak{R}}{e}{\left[{C}\,{\cdot}\,{\left({1} + {\gamma}\,{\cdot}\,{m}{\left({t}\right)}\right)}\,{\cdot}\,{Ae}^{{i}{2}{\pi}{f}_{c}{t}} \right]} \tag{6} \] meaning that the information is only accessible at the carrier frequency ${f}_{c}$. A single I/Q demodulation scheme (see Figure 1, upper part only) is sufficient to reconstruct ${m(t)}$. In the case of the PW mode, ${x(t)}$ is given by (4), and the information is now accessible at all of the carrier frequencies ${f}_{n}$, ${n}\,{\in}\,{\left\{{0},{\ldots},{N}{-}{1}\right\}}$. Namely, considering that the harmonic spacing ${f}_{{n} + {1}} - {f}_{n} = {\left({1} / {T}_{0}\right)}$ is sufficiently greater than the data rate of the tag [i.e., the frequency band occupied by ${m(t)}$], the signal received at the reader can be decomposed as \[{r}_{\text{PW}}{\left({t}\right)} = \mathop{\sum}\limits_{{n} = {0}}\limits^{{N}{-}{1}}{r}_{n}{(}{t}{)} = \mathop{\sum}\limits_{{n} = {0}}\limits^{{N}{-}{1}}{\frak{R}}{e}{\left[{C}_{n}\,{\cdot}\,{\left({1} + {\gamma}_{n}\,{\cdot}\,{m}{\left({t}\right)}\right)}\,{\cdot}\,{\alpha}_{n}{e}^{{i}{2}{\pi}{f}_{n}{t}}\right]}{.} \tag{7} \]
Figure 1. Block diagram of the DSP chain used to deliver ${\bigtriangleup}{V}_{{f}_{n}}{m}{\left({t}\right)}$ at each carrier frequency ${f}_{n}$. DSP: digital signal processing; LPF: low-pass filter; SRDS: sampling rate downsampler.
As a result, a coherent multicarrier I/Q demodulation scheme should be applied in PW mode to recover ${m(t)}$ in the most efficient way. This processing consists of applying an I/Q demodulation scheme (as in Figure 1) at each carrier frequency ${f}_{n}$ and summing the results. However, one must take care that the complex gains ${C}_{n}$ and ${\gamma}_{n}$ are frequency dependent. Because the accessible informative part of the signal received at the carrier frequencies ${f}_{n}$ is proportional to ${\frak{R}}{e}{\left[{\alpha}_{n}{C}_{n}{\gamma}_{n}\,{\cdot}\,{m}{\left({t}\right)}\right]}$, they are more than likely to sum up in an incoherent fashion, leading to poor communication performance. A complex domain rotation stage must therefore be added before the summation. Figure 1 illustrates the processing chain applied at each carrier frequency, eventually delivering the set of coherent signals ${m}_{n}{\left({t}\right)} = {\Delta}{V}_{{f}_{n}}{m}{(}{t}{)}$ that can be summed to recover the tag information.
Compared to the classical CW mode, where the carrier frequency ${f}_{c}$ is the only parameter that can be adapted to a given situation (within a certain bound as defined by the standard), the PW mode introduces a new degree of freedom to passive RFID systems. Namely, designing the pulse ${p(t)}$ or allocating its power across its ${N}$ carrier frequencies opens the possibility of further optimizing the performance of RFID systems. This flexibility is particularly interesting when the reader and the tag are separated by a complex multipath propagation channel, as will be discussed in the sections “PW RFID Adapted to Complex Multipath Propagation Channels” and “Simulation and Experimental Results.”
Several optimization criteria could be considered in the process of designing the pulse ${p(t)}$. A first criterion, which has already received thorough attention, is the maximization of the tag RF-to-dc rectifier circuit’s efficiency (see [7], [8], [9], and [10] to cite only a few sources). Those works have shown that better wireless power transfer performance is obtained by increasing the peak-to-average-power ratio (PAPR) of the signals incoming to the rectifier circuit. It is thus straightforward to optimize the shape of the pulse ${p}{\left({t}\right)}$ or, equivalently, to optimize its parameters ${\alpha}_{n}$ according to this criterion. Those results clearly pave the way toward PW mode RFID.
But considering the RFID system as a whole, a second optimization criterion that should be considered in the design of the pulse ${p(t)}$ is related to the information transfer in the tag-to-reader backward link. Because the electronic characteristics of the tag—as well as those of the propagation channel—are highly frequency dependent, the ability to recover the signal ${m(t)}$ can be drastically different from one carrier frequency to another. As a consequence, the possibility of designing the set of parameters ${\alpha}_{n}$ in such a way that the information recovery at the reader is optimized is another advantage brought by the PW mode.
In a multipath channel, the scattered electromagnetic wave splits into several paths. To mitigate the consequences of this type of channel, which is commonly encountered in practice, modern modulations, such as those based on orthogonal frequency division multiplexing, rely on the use of several carriers to exploit the frequency diversity. The strategy is to distribute the information across several frequency channels, such that even in the presence of strong frequency selectivity, a mean level of performance will be preserved. The concept of the PW mode in RFID is also based on this general principle, which consists of distributing the power on several carrier frequencies to avoid falling into a frequency hole. Figure 2 illustrates this principle by depicting the spectrum of the signal transmitted by the reader in CW mode [Figure 2(a)] and in uniform PW mode [Figure 2(b)]. As discussed in the next section, power allocation adapted to the channel [Figure 2(c)] will further increase the RFID system performance and is the basis of the time reversal (TR) mode.
Figure 2. Illustration of the options for distributing the transmitted power over one or more frequencies: (a) power carried by one frequency (CW mode), (b) power evenly distributed over 10 frequencies (uniform PW mode), and (c) power distributed over 10 frequencies according to channel selectivity (adapted PW mode).
The principle behind TR processing is to design signals that integrate the characteristics of the propagation channel in such a way that the propagating wave is focused spatially and temporally on a desired focal point. Without prior knowledge of the channel, the implementation of TR requires two stages, called the learning stage and the focusing stage [12], [13]. As illustrated by Figure 3(a) and (b), the learning stage consists of transmitting a short pulse signal ${u}{\left({t}\right)}$ and receiving the signal ${v}{\left({t}\right)}$. The received signal ${v}{\left({t}\right)}$ is composed of a succession of delayed and attenuated echoes, which fully characterize the multipath channel. In the focusing stage [Figure 3(c) and (d)], a time-reversed version of ${v}{\left({t}\right)}$ is transmitted, namely \[{v}_{\text{TR}}{\left({t}\right)} = {v}{\left({T}{-}{t}\right)} = {u}{\left({t}\right)}\,{*}\,{h}{\left({T}{-}{t}\right)} \tag{8} \]
Figure 3. Learning and focusing steps of the TR experiment assuming a Rayleigh channel model for the illustration.
where ${h}{\left({t}\right)}$ denotes the impulse response of the propagation channel, and ${T}$ is the duration considered. Note that ${u}{\left({t}\right)}$ and ${v}_{\text{TR}}{(t)}$ have been normalized for the illustration (Figure 3). The signal ${y}{\left({t}\right)}$ received after propagation is then focused on the receiver and exhibits a high PAPR as it has the highest possible peak power for a given power constraint on the emitted signal. TR processing is thus a way to design channel-adapted PW signals with high power transfer capacities.
In the context of PW mode RFID, the TR mode, as it will be called hereafter, is one option to obtain a channel-adapted PW mode based on the first criterion stated in the section “A New Degree of Freedom,” namely power transfer in the reader-to-tag forward link. The TR mode amounts to defining the pulse in (2) such that \[{p}_{\text{TR}}{\left({t}\right)} = {p}_{\text{UPW}}{\left({t}\right)}\,{\ast}\,{h}{(}{T}{-}{t}{)} \tag{9} \] where ${p}_{\text{UPW}}{(t)}$ is a pulse with a uniform distribution of power over its spectrum, as in Figure 2(b) (note that this pulse is only formally required to impose a finite bandwidth ${B}$ around the carrier frequency ${f}_{c}$) or, equivalently, to setting the coefficients ${\alpha}_{n}$ in (4) as \[{\alpha}_{n} = {H}^{*}{\left(\frac{n}{{T}_{0}}\right)} \tag{10} \] where ${H}^{\ast}{(}{f}{)}$ denotes the complex conjugate of the frequency response of the channel.
One should note that, for passive UHF RFID systems, the channel can be the reader-to-tag channel (including their antennas), but it also can be the full round-trip channel between the reader in emission mode and the reader in reception mode, integrating the forward link and the backward link as well as the tag’s backscattering capacities. However, in the first scenario (reader-to-tag channel only), it might be very difficult to get a priori knowledge of the channel at the reader side, unless the propagation channel is assumed to be invariant and can therefore be determined upstream and assumed stable, which is a strong assumption in most applications encountered in RFID. Some work has been proposed to acquire at least partial channel state information at the transmitter. But these methods require either integrating new processing capabilities at the tag level [14], [15] or using multiple antennas at the reader side [16]. In the second scenario, that is, when considering the full round trip, the channel is learned at the reader side from the signal received r(t), the processing is carried out entirely at the reader side, and the tag is implicitly integrated into the channel. The advantage of this solution is that it does not require changes to the tags that are already deployed in applications, but only to the readers.
The pulse waveform transmitted by the reader can be optimized on the second criterion mentioned in the section “A New Degree of Freedom,” namely the information transfer in the tag-to-reader backward link. As discussed in the section “Additional Processing at the Reader,” the signal ${r(t)}$ backscattered by the tag and received at the reader is a two-state modulated signal. The distance between those two states, accessible via the complex parameter ${\gamma}$ in (6), constitutes a noise robustness and separation ease criterion to distinguish between the two states of the tag. After processing of the received signal ${r(t)}$ at each carrier frequency, this separation capacity is given by the parameter ${V}_{{f}_{n}}$ introduced in the section “Additional Processing at the Reader.” A natural information-related strategy for the design of the pulse ${p(t)}$ is to allocate an amount of power proportional to ${\Delta}{V}_{{f}_{n}}$ to the carrier frequency ${f}_{n}$. The ${\Delta}{V}$-adapted PW mode, known as the adaptive power allocation (APA) mode hereafter, is thus defined by setting the coefficients ${\alpha}_{n}$ in (4) to \[{\alpha}_{n} = {\Delta}{V}_{{f}_{n}}{.} \tag{11} \]
In [17], the simplified UHF RFID system model shown in Figure 4 is implemented using the Ansys Electronics Desktop unified simulation platform for electromagnetic, circuit, and system simulation.
Figure 4. A schematic depiction of the simplified model of the passive UHF RFID system (note that the DSP block is detailed in Figure 1). AWG: arbitrary waveform generator.
Considering any type of channel, this simulation framework allows the comparison of the performance of several “strategies” to design ${p}{\left({t}\right)}$ in terms of energy transfer (i.e., dc power recovered by the tag) and backscattered information transfer (i.e., ${\Delta}{V}$ amplitude received by the reader). From three arbitrary propagation channels [free space (FS), LOS, and NLOS cases], Figure 5 compares the achieved results for the four modes: CW (uniform), PW, TR, and APA. For the APA mode, the power of all carriers ${f}_{n}$ is allocated proportionally to ${\Delta}{V}_{{f}_{n}}$. The power distribution on several frequencies globally improves the reader-to-tag energy transfer and the tag-to-reader information transfer, and this is all the more important as the channel properties are taken into account. The gain in the FS channel on the information transfer seems marginal, but for the CW mode, only the optimal case (i.e., the best of the frequencies) is shown. Finally, it should be noted that these results were obtained for arbitrarily chosen channels. If the trend observed here confirms expectations, the values obtained will remain relative (here the transmitted power was 15 dBm), and for some specific multipath channels the gain compared to CW may be more or less important.
Figure 5. Performance of CW, PW, TR, and APA modes in terms of (a) the harvested dc power (in microwatts) and (b) the resultant ${\bigtriangleup}{V}_{\Sigma}$ (in millivolts).
Commercial test and measurement platforms have been developed to meet the needs of RFID technology evaluation, for example, solutions such as Tagformance from Voyantic, RFID Xplorer from CISC Semiconductor, and VISN-100 from National Instruments. These platforms generally integrate one or more of the RFID standards, from low frequency to UHF, including near-field communication. They are therefore very useful for testing or verifying the performance of current RFID systems, and they are widely used in industrial and academic environments.
However, their use is increasingly limited by the evolution of RFID and its new applications. Several academic teams have proposed specific platforms that allow more flexibility or specific measures. Software-defined radio (SDR) offers a solution for emulating readers that easily achieves different configurations [18] as well as field-programmable gate array cards associated with front-end modules, which can also emulate tags [19]. This is why the SDR-based RFID reader approach is becoming increasingly popular; it allows for relatively low-cost solutions that are capable of integrating multistandard operations and new functionalities through easy-to-implement software upgrades, and thus it offers a high degree of flexibility as well [18], [20]. Concerning more specifically the study of optimized waveforms, there are few dedicated platforms. Some examples exist in the literature [21], [22], [23]; however, they remain limited in terms of flexibility because they do not allow arbitrary waveforms or unconventional modulation modes. Another limitation of this approach is that data are often postprocessed.
The platform RFID Waveformer can emulate an RFID reader in bistatic mode that is capable of transmitting nonconventional arbitrary waveforms, following the ISO 18000 protocol, and it can receive the signal back-modulated by the tag and demodulate it in real time [24]. The RFID Waveformer is presented in Figure 6. The system architecture consists of three main parts: the control interface, the RF transmitting and receiving equipment, and the propagation channel based on a reverberation box with metallic obstacles. To illustrate the potential of the RFID Waveformer, the three operating modes, CW (uniform), PW, and TR, are experimentally compared in terms of energy and information transfer. The three commercial tags shown in Figure 7 were chosen arbitrarily.
Figure 6. The RFID Waveformer. (a) The architecture system. (b) The experimental platform. (c) The reverberation channel and the barriers placed inside it. Rx: receiver; Tx: transmitter.
Figure 7. The three passive UHF RFID tags used for the measurements.
The emulated reader interrogates the tag placed in the channel using the standard ISO 18000 communication protocol in UHF RFID. The energy transfer from the reader to the tag is evaluated during phase 1 (i.e., the CW and query phases), while the information transfer from the tag to the reader is evaluated during phase 2 [i.e., when the tag sends back the 16-bit random sequence (RN16)]. Figure 8 illustrates the shape of the signal emitted by the emulated reader for low and high levels for each case.
Figure 8. The time domain of the transmitted signals for the two information bits in three modes: CW, PW, and TR.
This signal is written as \[{s}_{{\text{mod }}{e}}{\left({t}\right)} = {x}_{{\text{mod }}{e}}{(}{t}{)}\,{\cdot}\,{e}{\left({t}\right)} \tag{12} \] with the index “mode” indicating the mode (CW, TW, or TR). ${e}{\left({t}\right)}$ is the envelope of the transmitted signal, which is a constant signal during the remote feeding phase and a binary signal (the “query” information) during the interrogation phase (Figure 9).
Figure 9. The envelope of the signal transmitted by the reader when it starts the interrogation of the tag.
For the CW mode, ${x}_{{\text{mod }}{e}}{(t)}$ is a monochromatic signal [see (1)], while for the PW and TR modes, this signal consists of a train of pulses with a period ${T}_{0}$. For this experimental study, ${p}{\left({t}\right)}$ is predefined as a cardinal sine function, with finite support via Hamming windowing, which is modulated at the frequency${f}_{c}$, and whose spectral content is parameterized by the factor ${B}$ (note that the experimental results presented next are obtained for ${B} = {100}$, 200, and ${300}\,{MHz}$). From (2), ${x}_{\text{PW}}{\left({t}\right)}$ is therefore written as
and ${x}_{\text{TR}}{\left({t}\right)}$ is given by \[{x}_{\text{TR}}{\left({t}\right)} = {\left\{{p}{\left({t}\right)}\,{*}\,{h}{\left({T}{-}{t}\right)}\right\}}{\ast}\,{\text{III}}_{T_0}{(t)} \tag{14} \]
To evaluate the energy transfer, a receiving antenna is placed near the tag and can estimate the minimum received signal power necessary to activate the tag during phase 1. For a given frequency-selective channel, Table 1 summarizes the results obtained for the three tags by comparing the power gain (in dBm) obtained in the TR mode with the other two modes, PW and CW. It should be noted that, in the frequency band considered, the frequency chosen for the CW mode is the one that is optimal (i.e., the one that allows activation with minimum power). Overall, the TR mode uses less activation power than both the CW and the PW modes. However, the results obtained are very dependent on the tags, whose own bandwidth plays an important role.
Table 1. Tags’ activation power gains obtained with the TR mode compared to the PW mode (left) and the CW mode (right).
The evaluation of the transfer of information from the tag to the reader is based on the same experimental configuration as before but is carried out during phase 2 by comparing the difference ${\bigtriangleup}{V}$ between the two levels of information “0” and “1.” For the three tags considered, Table 2 shows the results obtained. For the CW case, the minimum and maximum values of ${\bigtriangleup}{V}$ (according to the set of carrier frequencies) are given; for the PW and TR cases, the ${\bigtriangleup}{V}$’s are obtained by exploiting all of the carrier frequencies, as described in the section “Additional Processing at the Reader.”
Table 2. Voltage difference DV between the low and high levels of the signal backscattered by the tag.
The results in the PW and the TR modes are quite close. As expected, the increase in the frequency band is to the advantage of the TR mode, which exploits each carrier more efficiently—unlike the PW mode, which considers all of the carriers in a uniform way. Compared to the optimal case of the CW mode, the PW and TR modes are not necessarily more efficient; for example, in the case of the SML GB4U8 tag, the CW mode appears to be the most efficient. However, even if, as in the previous case, these results depend on the tag, they are also and above all dependent on the channel. The optimal frequency to choose for the CW mode will not be known in advance, and, as the worst case shows, the voltage difference can become very small. Thus, PW and TR modes will be more robust, and, in particular, the PW mode seems to be the most efficient, with a performance similar to that of the TR mode but without requiring a channel learning phase.
Passive UHF RFID in PW mode relies on a certain frequency diversity to adapt to any type of propagation channel and thus to various application contexts. In itself, it is not necessarily more efficient than the current standard based on a single carrier frequency, but it becomes significantly more robust when the application is located in a complex propagation channel, especially one that is highly frequency selective. The solutions proposed here are designed to preserve the advantages of tags, notably their relatively low cost, their simplicity, and, above all, their passive nature as well as to take into account existing standards in terms of frequency band and transmitted power level. The increase in complexity is confined to the reader side, in transmission and in reception, with advantages that include the ability to transmit arbitrary waveforms and to receive information from multiple carriers, all associated with new processing capabilities.
With the RFID market continuing to grow at a double-digit annual rate and with the emergence of artificial intelligence, the current trend for manufacturers, especially RFID reader manufacturers, is to exploit the current UHF standard to its fullest potential and expand its applications, before considering new design paradigms that would require more expensive RFID readers with new hardware capabilities. Nevertheless, PW passive RFID is a very promising way forward. It not only improves the current performance in harsh environments but is also a source of inspiration for considering new applications: for example, the focusing properties offered by the RT application may allow increased security for the interrogation of a tag and/or the addition of localization capabilities.
[1] A. Goldsmith, Wireless Communications. Cambridge, U.K.: Cambridge Univ. Press, 2005.
[2] F. Perez Fontan and P. Marino Espineira, Modeling the Wireless Propagation Channel: A Simulation Approach with MATLAB®. Chichester, U.K.: Wiley, 2008.
[3] D. Kim, M. Ingram, and W. Smith, “Measurements of small-scale fading and path loss for long range RF tags,” IEEE Trans. Antennas Propag., vol. 51, no. 8, pp. 1740–1749, Aug. 2003, doi: 10.1109/TAP.2003.814752.
[4] J. D. Griffin and G. D. Durgin, “Complete link budgets for backscatter-radio and RFID systems,” IEEE Antennas Propag. Mag., vol. 51, no. 2, pp. 11–25, Apr. 2009, doi: 10.1109/MAP.2009.5162013.
[5] M. Buettner and D. Wetherall, “An empirical study of UHF RFID performance,” in Proc. 14th ACM Int. Conf. Mobile Comput. Netw., San Francisco, CA, USA, Sep. 2008, pp. 223–234, doi: 10.1145/1409944.1409970.
[6] M. Bertocco, A. Dalla Chiara, G. Gamba, and A. Sona, “Experimental analysis of UHF RFID impairments and performance,” in Proc. IEEE Instrum. Meas. Technol. Conf., Singapore, May 2009, pp. 759–764, doi: 10.1109/IMTC.2009.5168552.
[7] A. Boaventura, D. Belo, R. Fernandes, A. Collado, A. Georgiadis, and N. B. Carvalho, “Boosting the efficiency: Unconventional waveform design for efficient wireless power transfer,” IEEE Microw. Mag., vol. 16, no. 3, pp. 87–96, Apr. 2015, doi: 10.1109/MMM.2014.2388332.
[8] B. Clerckx and E. Bayguzina, “Low-complexity adaptive multisine waveform design for wireless power transfer,” IEEE Antennas Wireless Propag. Lett., vol. 16, pp. 2207–2210, May 2017, doi: 10.1109/LAWP.2017.2706944.
[9] M. S. Trotter and G. D. Durgin, “Survey of range improvement of commercial RFID tags with power optimized waveforms,” in Proc. IEEE Int. Conf. RFID, Orlando, FL, USA, Apr. 2010, pp. 195–202, doi: 10.1109/RFID.2010.5467265.
[10] A. J. S. Boaventura and N. Carvalho, “Extending reading range of commercial RFID readers,” IEEE Trans. Microw. Theory Techn., vol. 61, no. 1, pp. 633–640, Jan. 2013, doi: 10.1109/TMTT.2012.2229288.
[11] P. V. Nikitin, K. V. S. Rao, and R. Martinez, “Differential RCS of RFID tag,” Electron. Lett., vol. 43, no. 8, pp. 431–432, Apr. 2007, doi: 10.1049/el:20070253.
[12] G. Lerosey, J. de Rosny, A. Tourin, A. Derode, G. Montaldo, and M. Fink, “Time reversal of electromagnetic waves,” Phys. Rev. Lett., vol. 92, no. 19, May 2004, Art. no. 193904, doi: 10.1103/PhysRevLett.92.193904.
[13] M. Zhang, C. Fang, P. Doanis, J. Huillery, A. Bréard, and Y. Duroc, “Time-reversal processing for downlink-limited passive UHF RFID in pulsed wave mode,” IEEE Antennas Wireless Propag. Lett., vol. 18, no. 12, pp. 2562–2566, Dec. 2019, doi: 10.1109/LAWP.2019.2943211.
[14] Y. Huang and B. Clerckx, “Waveform design for wireless power transfer with limited feedback,” IEEE Trans. Wireless Commun., vol. 17, no. 1, pp. 415–429, Jan. 2018, doi: 10.1109/TWC.2017.2767578.
[15] S. Shen, J. Kim, C. Song, and B. Clerckx, “Wireless power transfer with distributed antennas: System design, prototype, and experiments,” IEEE Trans. Ind. Electron., vol. 68, no. 11, pp. 10,868–10,878, Nov. 2021, doi: 10.1109/TIE.2020.3036238.
[16] G. Yang, C. K. Ho, and Y. L. Guan, “Multi-antenna wireless energy transfer for backscatter communication systems,” IEEE J. Sel. Areas Commun., vol. 33, no. 12, pp. 2974–2987, Dec. 2015, doi: 10.1109/JSAC.2015.2481258.
[17] H. Ezzedine, Y. Merakeb, J. Huillery, A. Bréard, and Y. Duroc, “Simulation framework for studying UHF RFID systems in pulse wave mode,” in Proc. IEEE Int. Conf. RFID Technol. Appl. (RFID-TA), 2019, pp. 120–124, doi: 10.1109/RFID-TA.2019.8892067.
[18] L. Catarinucci, D. De Donno, R. Colella, F. Ricciato, and L. Tarricone, “A cost-effective SDR platform for performance characterization of RFID tags,” IEEE Trans. Instrum. Meas., vol. 61, no. 4, pp. 903–911, Apr. 2012, doi: 10.1109/TIM.2011.2174899.
[19] C. Angerer and R. Langwieser, “Flexible evaluation of RFID system parameters using rapid prototyping,” in Proc. IEEE Int. Conf. RFID, Orlando, FL, USA, Apr. 2009, pp. 42–47, doi: 10.1109/RFID.2009.4911188.
[20] G. Saxl, L. Goertschacher, T. Ussmueller, and J. Grosinger, “Software-defined RFID readers: Wireless reader testbeds exploiting software-defined radios for enhancements in UHF RFID Systems,” IEEE Microw. Mag., vol. 22, no. 3, pp. 46–56, Mar. 2021, doi: 10.1109/MMM.2020.3042408.
[21] E. A. Keehr and G. Lasser, “Making a low-cost software-defined UHF RFID reader,” IEEE Microw. Mag., vol. 22, no. 3, pp. 25–45, Mar. 2021, doi: 10.1109/MMM.2020.3042046.
[22] M. S. Trotter, J. D. Griffin, and G. D. Durgin, “Power-optimized waveforms for improving the range and reliability of RFID systems,” in Proc. IEEE Int. Conf. RFID, Orlando, FL, USA, Apr. 2009, pp. 80–87, doi: 10.1109/RFID.2009.4911196.
[23] A. J. S. Boaventura and N. B. Carvalho, “The design of a high-performance multisine RFID reader,” IEEE Trans. Microw. Theory Techn., vol. 65, no. 9, pp. 3389–3400, Sep. 2017, doi: 10.1109/TMTT.2017.2663405.
[24] Y. Merakeb, H. Ezzedine, J. Huillery, A. Bréard, R. Touhami, and Y. Duroc, “Experimental platform for waveform optimization in passive UHF RFID systems,” Int. J. RF Microw. Computed-Aided Eng., vol. 30, no. 10, Oct. 2020, Art. no. e22376, doi: 10.1002/mmce.22376.
Digital Object Identifier 10.1109/MMM.2023.3293613