Martin Maderböck, Thomas Ussmueller
©SHUTTERSTOCK.COM/PEERA_STOCKFOTO
With the increasing popularity of industry 4.0 applications and the integration of network-capable small devices into the Internet of Things (IoT), the technical requirements for the transmission of information are also rising. High reliability is needed, especially for wireless systems used for industrial applications. One of the biggest challenges in this context is multipath propagation. Ultrahigh frequency radio-frequency identification, or UHF RFID for short, has proven to be a reliable technology for reading and transmitting data. In this communication system, the data records to be transmitted are modulated by a transmitter (reader) to an electromagnetic alternating field with a frequency between 860 MHz and 960 MHz and serve the receivers (transponders) not only for information transmission but also for energy supply.
This is particularly critical in the case of battery-less (passive) transponders, since the limiting factor in the transmission range is the transmission of the required energy. This parameter, known as power sensitivity, indicates the lower limit of transmitted power required to operate the RFID chip on the transponder. It is heavily dependent on a transponder’s antenna and on circuits such as the power rectifier and limiter [2], [3], [4].
The lower the sensitivity of a transponder, the greater its range. An increase in transmitting power therefore also leads to a greater possible communication distance. However, power levels are limited by country-specific specifications. In North America, for example, a maximum transmission power of the reader of 4 W is permitted [5]. For Europe, the maximum transmission power is limited to ${ERP}_{\max} = {2}{\text{ W}}$, in this case, and the received power with respect to distance is depicted in Figure 1.
Figure 1. Received power ${P}_{T}$ at a transponder according to the Friis transmission equation for 865 MHz with a transmitted power of 2 W [1].
This power limit can be circumvented by introducing an external signal source. When it comes to improving RF links by means of multiple carriers, two different approaches exist as depicted in Figure 2(b) and (c). The first approach, which was first mentioned by Nikitin et al. in 2006, involves adding one or more continuous wave (CW) emitters to an existing RFID reader in order to supply the tag with supplementary power from the additional emitters [7]. The CW signal can either be precisely frequency and phase locked to the reader’s signal in order to enhance the emitted wave, or in a multicarrier system, independent carriers with different frequencies are employed. To allow a seamless demodulation of the data, the frequency difference between carriers must be large enough so that the resulting beat frequency can be separated from the modulated data by the tag’s internal low-pass filter. The second approach, which was developed by Trotter in 2009, uses multiple carriers to create so-called power optimized waveforms [8]. This principle relies on the typically used diode type rectifiers in a transponder. The beat signal created by multiple carriers with different frequencies improves the efficiency of the charge pump circuit and therefore allows longer read ranges, with the same total power emitted. In this article, a novel approach is presented, which uses two carrier waves with a large frequency separation in order to counter fading in a multipath environment. Additionally, by using two carrier waves for communication with the tag, immunity to narrowband disturbance and interference can be increased [6].
Figure 2. Different approaches of reader to tag link. (a) shows the typical communication scheme, where one carrier frequency is used for communication. In (b), a reader plus CW approach is sketched. A reader sends the amplitude shift keying modulated commands on one frequency. Additionally, CW emitters are placed in closer vicinity to the tag in order to supply it with extra power, often using a different frequency. (c) shows the multicarrier concept presented in this article. In that case, one reader sends a “query” on two frequencies simultaneously and evaluates the tag’s answer on both channels to gain additional redundancy [6].
Systems that use an external, unmodulated carrier wave also offer better transponder reachability, due to compensation for dead spots. These are directly related to the transmission frequency and form the system environment due to wave superposition of multipath propagation [9]. Thus, additional energy supply introduced by external carrier wave ensures the coverage of vulnerable system points [10].
The introduction of such external energy supply into the system is not arbitrarily scalable. Just as several additional unmodulated carrier waves compensate for dead spots, however, they harm the modulated data signal. The modulation depth of an amplitude-modulated signal is in danger of being reduced as a result. If this falls below a certain threshold, the logic of the transponder is not able to interpret the data correctly [11]. Thus, communication fails not because of the power supply, but because of the loss of information in the data signal.
By means of the backscatter modulation method, the transponder changes its reflection properties by means of impedance tuning of the antenna, resulting in a modulated reflection of the incident carrier wave. This process makes it possible to transmit the data to the reader with extremely low energy consumption [12]. In the case of additional unmodulated carrier waves introduced into the system, this causes a response to each of these transmitters which, do not detect them. If, however, modulated data signals are involved, which have different frequencies and are synchronized with each other, this becomes a full-fledged multicarrier process. Essential for this is a reader suitable for multicarriers, which evaluates the transponder responses on all transmission frequencies. This redundant information transmission also serves to ensure stable communication through error correction.
In multicarrier systems, care must be taken to ensure that the transmission frequencies used are sufficiently different from one another. If sufficient spacing is not selected, overlapping will result in a distorted data signal that the transponder cannot interpret [13]. In addition, the interference pattern is very similar. Frequency-dependent dead spots are thus close to each other or even overlap.
The regulations in force in Europe for the commercial use of UHF RFID systems describe the short-range device band, which covers a frequency range from 863 MHz to 870 MHz. This restriction has so far prevented the practical, industrial use of a multicarrier process. With plans for European harmonization of the North American industrial, scientific, and medical band, which frees the frequency range from 902 MHz to 928 MHz, the multicarrier process is gaining renewed attention [14].
RFID systems are often used in multipath environments, where the fading effects can lead to a loss of signal. To counter this, different approaches exist which rely on antenna diversity using spatially distributed antennas within the region of interest, resulting in a strongly different fading pattern for each antenna. An interrogator can either switch between the different patterns (antennas) or combine them in a beneficial way. Another way of creating diversity to counter fading effects is achieved by frequency hopping. The necessary bandwidth to achieve a different fading pattern depends on properties of the environment as well as the transmission range. In [15], a complete system is demonstrated, which employs an antenna array together with a frequency and phase hopping scheme. The system randomly switches between different antennas, frequencies, and phases in a timely manner. Thereby, a significant improvement of coverage in a test room can be achieved, compared to a traditional system. However, these systems do not increase the success rate of a single command, but they improve coverage by changing the readable area over time [6].
Basic considerations of multicarrier methods in a real measurement environment in [6] show the advantages of the method for UHF RFID. An increased reception coverage of the RFID transponder by correct response along a 1D measurement could be demonstrated with it. Furthermore, a method is described which uses the generated redundant information for bitwise error correction. These initial results form the basis of this work for further consideration and characterization of the radio channel for multicarrier applications.
The term dead spot plays a major role in any transmitter–receiver system, including UHF RFID communications. While too great a distance between transmitter and receiver is the most common cause of communication interruption (transmitted power too low), spatial conditions can also lead to this phenomenon. To describe the effect with electromagnetic waves, the physical term interference is necessary.
In general, interference is understood to be the change in amplitude when two or more waves are superimposed and added together with the correct sign. It is presupposed that the waves to be added have the same frequency and a similar amplitude. If these requirements are met, superposition under the correct boundary conditions leads either to an increase in amplitude (constructive interference) or to complete cancelation of the waves (destructive interference). The decisive factor for the result of the superposition is the phase shift between the added waves, also called path difference. A maximum increase in amplitude occurs when the waves are shifted by ${360}^{\circ}{n} = {2}{\pi}{n}$ (applies to ${n}\,{\in}\,{\Bbb{Z}}$), i.e., are in phase. Destructive interference, on the other hand, occurs when two waves shifted by ${180}^{\circ}{(}{2}{n} + {1}{)} = {\pi}{(}{2}{n} + {1}{)}$ (holds for ${n}\,{\in}\,{\Bbb{Z}}$) are added.
Superposition of waves occurs in almost every real radio system, due to multipath propagation by reflection. This leads in special cases to disturbing interferences, which mean a weakening of the energy transmitted to the receiver [16]. If the UHF RFID transponder cannot generate energy from the weakened carrier wave or from the carrier wave that is no longer present due to cancelation at a point in space, the result is an interruption in communication. This point of the system is a dead spot.
To make a valid prediction of the transmitted power when multipath propagation occurs, [17] formulates an equation for calculating multipath loss ${L}_{\text{path}}$. This allows multiple reflections to be considered simultaneously by summing all individual reflections as follows: \[{L}_{\text{path}} = {\left(\frac{\lambda}{{4}{\pi}{s}_{0}}\right)}^{2}{\left\vert{1} + \mathop{\sum}\limits_{{n} = {1}}\limits^{N}{\Gamma}_{n} \frac{{s}_{0}}{{s}_{n}}{e}^{{-}{jk}{(}{s}_{n}{-}{s}_{0}{)}}\right\vert}^{2}\,{\text{ with }}{k} = \frac{{2}{\pi}}{\lambda}{.} \tag{1} \]
In (1), the distance ${s}_{0}$ describes the direct radio link of the transmitting and receiving antennas, N the total number of reflections considered, ${s}_{n}$ the ${n}^{\text{th}}$ reflection path, and ${\Gamma}_{n}$ the Fresnel-coefficient of the n-reflection.
Thus, the multipath attenuation ${L}_{\text{path}}$ is to be used in the Friis transfer equation. If no reflections occur, the calculations simplify to the equations of Friis transmission as follows: \[\frac{{P}_{{r}_{\text{multi}}}}{{P}_{t}} = {G}_{t}\,{\cdot}\,{G}_{r}\,{\cdot}\,{\left(\frac{\lambda}{{4}{\pi}{s}_{0}}\right)}^{2}\,{\cdot}\,{\left\vert{1} + \mathop{\sum}\limits_{{n} = {1}}\limits^{N}{{\Gamma}_{n}} \frac{{s}_{0}}{{s}_{n}}{e}^{{-}{jk}{(}{s}_{n}{-}{s}_{0}{)}}\right\vert}^{2}{.} \tag{2} \]
To calculate in decibels, convert the term and write (4) as follows: \begin{align*}{P}_{{r}_{\text{multi (dB)}}} = & {P}_{{t}{\text{(dB)}}} + {G}_{{t}{\text{(dB)}}} + {G}_{{r}{\text{(dB)}}} + {L}_{\text{path (dB)}} \tag{3} \\ {L}_{\text{Path (dB)}} = & {20}\,{\cdot}\,{\log}_{10}{\left(\frac{\lambda}{{4}{\pi}{s_0}} \right)} \\ & + {20}\,{\cdot}\,{\log}_{10}{\left({\left\vert{1} + \mathop{\sum}\limits_{n = 1}^{N} {\Gamma}_{n} \frac{s_0}{s_n}{e}^{-jk{\left({s_n} - {s_0} \right)}}\right\vert}\right)}. \tag{4} \end{align*}
Together with the calculation of reflection paths, this model forms an effective view of multipath propagation. However, it only captures single reflections. This is sufficient for a first statement about the measurement environment because the free space attenuation influences these multipaths even more and therefore it changes the result only slightly.
Figure 3 shows the effect of multipath propagation on the amplitude of the transmitted power. Due to the advantage of expressing power in decibels, the multipath attenuation ${L}_{\text{path (dB)}}$ represents the actual power response of the measurement environment, since it is the only path-dependent summation term in (4). The free-space attenuation ${1} / {D}_{{f}{\text{(dB)}}}$, on the other hand, represents the theoretical power of the measurement area without reflections. It shows a dip of the transmit power at the point of maximum destructive superposition of the interference pattern ${I}{(}{\lambda},{s}_{0},{s}_{1}{)}$.
Figure 3. Example power curve of theoretical free-space attenuation ${1} / {D}_{{f}{\text{(dB)}}}$ without interference and multipath attenuation ${L}_{\text{path (dB)}}$ with single reflection present on drawn wall as well as an example interference curve ${I}{(}{\lambda},{s}_{0},{s}_{1}{)}$. This curve is determined by the path difference for the measuring space that begins at the initial position ${[}{x}_{{R}_{A}},\,{y}_{{R}_{A}}{]}$ and ends at the final position ${[}{x}_{{R}_{E}},{y}_{{R}_{E}}{]}$. At points where a phase shift of ${180}^{\circ}$ occurs, the interference is at its maximum destructive level. Conversely, when—due to multipath propagation—there is a phase shift of ${0}^{\circ}$ or ${360}^{\circ}$, the interference is at its maximum constructive level. Therefore, the attenuation ${L}_{\text{path (dB)}}$ takes both space attenuation and attenuation due to multipath interference into account [18].
The European Commission has recently decided to harmonize the frequencies for RFID, so that members have been called upon implementing new bands in the 915 MHz to 921 MHz range in early 2019. The maximum allowed radiated power is 2 W in the lower bands and up to 4 W in the upper bands. The reader is allowed to operate in both bands simultaneously. Therefore, in multicarrier UHF RFID setups the frequencies 868 MHz and 915 MHz are used. [6]
A suitable measurement setup was implemented for automatic characterization of the radio channel in any room. [18] This can be seen in Figure 4. For recording spatial phenomena, a 2D measurement is possible due to the setup. The total measurement range is about 3160 × 1060 mm. With a size of 20 mm, the grid is fine-meshed (${\approx}\,{1} / {17}$ of wavelength) enough to detect power dips. Smaller grid sizes do not provide any added value, which puts the significantly increasing time required for the measurement into perspective.
Figure 4. Complete measuring setup—with transmitting (at position ${[}{x}_{T},{y}_{T}{]}$) and receiving (at position ${[}{x}_{R},{y}_{R}{]}$) antenna drawn in, both base plates for maximum travel, measuring range drawn in, as well as spare picture of the required measuring equipment [18].
The UHF RFID reader system is based on the National Instruments LabVIEW-controlled PXIe platform, which consists of a PXIe-1075 Chassis and 18 possible slots as depicted in Figure 5. The slots are filled with various modules, including a controller module PXIe-8135 and a transceiver module PXIe-5644R. This transceiver operates in the frequency range of 65 MHz to 6 GHz with a bandwidth of 80 MHz. For the experiment to characterize the radio channel, only a CW source connected to the transmitting antenna was used. For the acquisition of data, an antenna (Siemens Simatic RF660A) is connected to the transceiver over a circulator to shield the outgoing and incoming signal. For the graphical evaluation of the data and the communication with the controller, a multitouch display (LG23ET83V) is used. The programming of the PXIe platform is based on the LabVIEW software and can be split into the transmitting and receiving part. The latter one takes the incoming signal from the antenna and displays it at the front panel. The following sections will not consider this straightforward process and instead focus on the transmitting part. The aim of the transmitting part is to generate an EPCglobal Gen2 V2 compliant Query-Command and transfer it to the antenna [19].
Figure 5. View of the experimental test setup depicting the reader device consisting of a National Instruments PXI, an amplifier, a circulator as well as the transmitting antenna [6].
Figure 6 shows the measurement environment. A room can be seen in which filing cabinets are located at different heights. This arrangement makes prediction difficult due to multipath propagation. The measurement area, the coordinate system, as well as the transmitting antenna and data cables are drawn in the image. An attached webcam allows the measurement process and progress to be observed without having to enter the room.
Figure 6. Measurement setup in real environment with occurring multipath propagation of the transmitted signals [18].
After a total measurement time of 15 h the experiment was completed. Figure 7 shows the result of the measurement for the two frequencies 868 MHz and 916 MHz. In each case, a plane is shown which represents a sensitivity attenuation of –28.55 dB. It represents the maximum attenuation, if the example Wireless Identification and Sensing Platform. Transponder with a sensitivity of 2.45 dBm is still operational at a power source of 31 dBm. All measurement points whose power value is below this level offer no possibility of receiving a response from the transponder.
Figure 7. Result of measurement in real multipath environment with transponder-sensitivity limit entered. Cable attenuation is included at a transmit power of 31 dBm [18].
The measurement curve in Figure 7 indicates by its ripple an interference pattern which is generated mainly by multiple occurring and overlapping reflections leading to strong amplitude fluctuations (deep dips).
Looking at the power distribution, it is obvious that measurement points located at a short distance from the transmitting antenna receive the greatest transmitted power. The interference phenomena increase significantly when approaching the walls of the measurement room, but not exclusively destructively. At a frequency of 868 MHz, a renewed increase of the transmitted power at the rear limit of the measurement range is clearly evident. However, severe power losses are also observed in the form of deep holes in points that are a short distance from the transmitting antenna [18].
The determination of the power in each measurement point is done separately for the two frequencies 868 MHz and 916 MHz. The effect of a multicarrier method is in the following determined theoretically.
When calculating the multicarrier behavior, it is important to note that an addition of the measured values in dBm does not correspond to a sum power in (milliwatts). Equation (5) gives the calculation rule: The power values are first to be converted into milliwatts, then added and are available in the unit dBm by forming the logarithm as follows: \[{P}_{\text{MC}}{(}{dBm}{)} = {10}\,{\cdot}\,{\log}_{10}{\left({10}^{\frac{{P}_{868{\text{MHz}}}{(dBm)}}{{10}{\text{ mW}}}} + {10}^{\frac{{P}_{916{\text{MHz}}}{(dBm)}}{{10}{\text{ mW}}}}\right)}{.} \tag{5} \]
Figure 8 represents the coverage of both frequencies, as well as the calculated multicarrier application. In addition, the percentage coverage provides information on how many of the measurement points receive sufficient energy, i.e., have a higher power than the transponder sensitivity [18].
Figure 8. Increasing the coverage by considering a multicarrier system. With included cable attenuation at a transmit power of 0 dBm [18].
As demonstrated by [6], multicarrier systems can also increase the coverage area by correcting bit errors. This works if these errors occur on both channels at different positions. This makes it possible to even correctly demodulate data if both channels can’t be demodulated on their own. This circumstance is shown in Figure 9.
Figure 9. Increasing the coverage with error correction using multicarrier systems [6].
Figure 8 clearly shows the coverage at different transmission frequencies, which differ significantly from each other. While the measurement at 916 MHz shows a larger coverage, at 868 MHz small islands (although larger distance) form where the transponder responds. While none of the single frequencies provides coverage greater than just under 40%, the theoretically applied multicarrier method has sufficient performance in 70% of the measurement points. This immense increase is due to positions of the transponder that just do not reach the sensitivity limit. The increase in power with superimposition thus creates new areas in which the reader receives a response from the transponder. Also, error detection algorithms are usable to enhance the coverage with the help of multicarrier systems. Furthermore, it has been shown that—given proper antenna matching—transponders can operate at 868 MHz and 915 MHz, respectively, without any changes.
[1] M. Ferdik, G. Saxl, and T. Ussmueller, “Battery-less UHF RFID controlled transistor switch for Internet of Things applications — A feasibility study,” in Proc. IEEE Topical Conf. Wireless Sens. Sens. Netw. (WiSNet), 2018, pp. 96–98, doi: 10.1109/WISNET.2018.8311574.
[2] D. Mair et al., “Numerical optimization of a fully cross-coupled rectifier circuit for wireless passive ultra low power sensor nodes,” Sensors, vol. 19, Oct. 2019, Art. no. 4527, doi: 10.3390/s19204527.
[3] D. Mair et al., “A time dependent temperature compensated limiter for passive differential UHF RFID,” in Proc. Inst. Elect. Electron. Eng., Jan. 2021, pp. 690–693, doi: 10.23919/EuMC48046.2021.9337941.
[4] D. Mair et al., “Evolutionary optimization of asymmetrical pixelated antennas employing shifted cross shaped elements for UHF RFID,” Electronics, vol. 9, no. 11, pp. 1–11, Nov. 2020, doi: 10.3390/electronics9111856.
[5] P. Nikitin and K. V. S. Rao, “Performance limitations of passive UHF RFID systems,” in Proc. IEEE Antennas Propag. Soc. Int. Symp., Jan. 2006, pp. 1011–1014, doi: 10.1109/APS.2006.1710704.
[6] M. Fischer et al., “An experimental study on the feasibility of a frequency diverse UHF RFID system,” IEEE Access, vol. 7, pp. 132,311–132,323, Sep. 2019, doi: 10.1109/ACCESS.2019.2939613.
[7] P. Nikitin, V. Kodukula, and F. Sander Lam, “Method and apparatus to increase the range of RFID systems,” U.S. Patent 11 558 339, Nov. 2006.
[8] 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 (RFID), 2010, pp. 195–202, doi: 10.1109/RFID.2010.5467265.
[9] Z. Hongsheng, “Analysis of dead zone by mathematical methods,” in Proc. IEEE 7th Int. Conf. Wireless Commun., Netw. Mobile Comput., 2011, pp. 1–4, doi: 10.1109/wicom.2011.6040569.
[10] P. V. Nikitin and K. V. S. Rao, “Antennas and propagation in UHF RFID systems,” in Proc. IEEE Int. Conf. RFID (Freq. Identif.), May 2008, pp. 277–288, doi: 10.1109/RFID.2008.4519368.
[11] H.-C. Liu and C.-R. Kuo, “Reader coverage analysis for multi-carrier passive UHF RFID systems,” in Proc. Wireless Pers. Commun., Jul. 2011, pp. 123–133, doi: 10.1007/s11277-010-0194-2.
[12] P. Sorells, “Passive RFID basics,” Microchip Technol. Inc., AN680, Chandler, AZ, USA, 1998. Accessed: May 8, 2021. [Online] . Available: http://ww1.microchip.com/downloads/en/appnotes/00680b.pdf
[13] H.-c. Liu, Y.-t. Chen, and W.-s. Tzeng, “A multi-carrier UHF passive RFID system,” in Proc. Int. Symp. Appl. Internet Workshops, 2007, pp. 21–21, doi: 10.1109/SAINT-W.2007.9.
[14] “Harmonised standard for access to radio spectrum,” ETSI, Sophia Antipolis, France, ETSI Standard Draft ETSI EN 302 208 V3.3.0, May 2020.
[15] S. Sabesan et al., “Wide area passive UHF RFID system using antenna diversity combined with phase and frequency hopping,” IEEE Trans. Antennas Propag., vol. 62, no. 2, pp. 878–888, Feb. 2014, doi: 10.1109/TAP.2013.2290114.
[16] K. Brocks, Die Messung Der Reflexionseigenschaften Künstlicher Und Natürlicher Materialien Mit Quasi-Optischen Methoden Bei Mikrowellen. 1st ed. Wiesbaden, Germany: VS Verlag für Sozialwissenschaften, 1957.
[17] M. Cremer et al., “An improved channel model for passive UHF RFID systems,” in Proc. IEEE Int. Conf. RFID-Technol. Appl. (RFID-TA), 2013, pp. 1–7, doi: 10.1109/RFID-TA.2013.6694511.
[18] J. Walk et al., “A multicarrier communication method to increase radio coverage for UHF RFID,” in Proc. 52nd Eur. Microw. Conf. (EuMC), 2022, pp. 872–875, doi: 10.23919/EuMC54642.2022.9924390.
[19] M. Ferdik et al., “NI PXIe based UHF RFID reader,” in Proc. 11th German Microw. Conf. (GeMiC), 2018, pp. 303–306, doi: 10.23919/GEMIC.2018.8335090.
Digital Object Identifier 10.1109/MMM.2023.3293636