3. Results
The presence of a metal vehicle cabin (modelled as a metal wall) and a person’s influence on the spatial distribution of EMF emitted by the HF RFID reader (especially in the surroundings of a bystander), as shown in
Figure 5 and
Figure 6 and
Table 1 and
Table 2, were analysed.
The electric and magnetic field values were evaluated in all the considered exposure scenarios and in the vicinity of RFID reader alone in the free space (i.e., in a numerical model with a reader on a vertical column, representing an unperturbed field distribution) or attached to the vehicle cabin, considering the output power of a reader sufficient to ensure the RR of 10 cm (according to ISO/IEC 14443-2:2020 requirements for PICCs of classes 1–3). The spatial distribution of the electric and magnetic fields was analysed in detail in nine points, as shown at
Figure 5a,e:
In a horizontal cross-section perpendicular to the reader plane (at a 120 cm height):
- –
A1 and A2—behind the person using the PICC device to their left and right side, respectively (mirror image to the centre of the reader), in the bystander’s location;
- –
B1 and B2—close to the left and right side of the person using the PICC device, respectively (mirror image to the centre of the reader), in the bystander’s location;
- –
C1 and C2—in front of the person using the PICC device to their left and right side, respectively (mirror image to the centre of the reader), in the bystander’s location;
Along a vertical line in the reader plane—D, E, F at heights of 170, 70 and 20 cm, respectively.
The highest variations of the point-values of the magnetic field in comparison to the unperturbed magnetic field values (from the model of the reader in a free space) were found in exposure scenarios with the reader attached to the vehicle cabin: an increase up to 40% (points C1 and C2) and a decrease up to 30% (points A1, B1, A2 and B2)—for points in the horizontal cross-section perpendicular to the centre of the reader plane, and an increase up to 50%—for points along the vertical line in the reader plane (
Table 1). Corresponding to these, the increase in exposure scenarios with a reader on a vertical column were up to 15% and 20%, respectively. The increase (points A1 and B1) or decrease (point A2) related to the presence of a bystander in the modelled scenario did not exceed 10%.
A greater variation of the point values of the electric field than the magnetic field was found. The highest variations in the values of the electric field, in comparison to unperturbed electric field values (from the model of the reader in a free space), were found (contrary to the magnetic field) in exposure scenarios with the reader on a vertical column. Up to a 250% increase (points A1 and B1) and up to a 20% decrease (points C1, A2, B2 and C2)—for points in the horizontal cross-section perpendicular to the reader plane, and up to a 100% increase—for points along a vertical line in the reader plane—were found for them (
Table 2). In exposure scenarios with the reader attached to the vehicle cabin, up to a 120% increase (points A1 and B1) and up to a 60% decrease (points C1, A2, B2, C2, D, E and F) were observed, respectively. The increase related to the presence of a bystander in the modelled scenario did not exceed 170%.
It should be noted that the increase in the reading range is related to a significant increase in the magnetic and electric field strengths at any point in the vicinity of the HF RFID reader. Comparing the investigated HF RFID reader (with inner dimensions of 80 × 80 mm and RR = 10 cm, corresponding to electric and magnetic field levels in its vicinity, as presented in
Table 1 and
Table 2), the values of the magnetic and electric field strength at particular locations were found to be lower for shorter RRs (e.g., 80% for an RR = 4 cm and 40% for an RR = 8 cm) and higher for longer RRs (e.g., 60%, 240% and 510% for RR of 12, 16 and 20 cm, respectively).
The normalised distributions of SAR and Ein values in the human bodies exposed to EMF near an HF RFID reader in various exposure scenarios (side view of values on the surface of the human body and values in the body horizontal cross-section perpendicular to the reader plane) are shown in
Figure 7 and
Figure 8.
Table 3 and
Table 4 show the results of the numerical simulations of SAR and Ein values, respectively, related to exposure to EMF at 13.56 MHz near an HF RFID reader with an antenna with inner dimensions of 8 × 8 cm operating with an output power sufficient to ensure an RR of 10 cm (according to ISO/IEC 14443-2:2020 requirements for PICCs of classes 1–3). Results were obtained with respect to continuous exposure (typical for validators) and worse-case exposure, associated with respect to the exposure duration considered in the SAR averaging time required by international guidelines and standards (6 min for local SAR10g and 30 min for WBSAR). It is also worse-case exposure when Ein is considered without time-averaging.
The highest SAR values were found in the model of a person using a PICC device in an exposure scenario with two people and an HF RFID reader (validator) located on a vertical column—7.4 mW/kg for SAR10g in head and torso, 46 mW/kg for SAR10g in limbs and 0.28 mW/kg for WBSAR (
Table 3). No significant differences between SAR values in the model of a person using a PICC device obtained for all the investigated exposure scenarios were found—all observed differences were below the level of uncertainty for the carried out numerical simulations, estimated as ±25% (K = 1). The highest differences of 5% were found between values obtained for exposure scenarios with various HF RFID reader locations (on a vertical column and attached to the metal vehicle cabin). In the case of a bystander, the highest SAR values were found in the same exposure scenario where the highest SAR values were found for the person validating a public transport card—0.07 mW/kg for SAR10g in the head and torso, 0.06 mW/kg for SAR10g in limbs and 0.002 mW/kg for WBSAR. These values were at least 100 times lower than the SAR values mentioned above, obtained in the model of a person using a PICC device, and up to 2 times higher than SAR values obtained for an HF RFID reader attached to the metal vehicle cabin.
All other SAR values in both human body models (of a person using a PICC device and the bystander) were below 1.2% of the limits of the general public (GP) exposure (ICNIRP 2020)/exposure in unrestricted environments (IEEE), and thus the limits of occupational exposure (OE) (ICNIRP)/exposure in restricted environments (IEEE).
The highest peak Ein values were found in the model of a person using a PICC device and were found also in the same exposure scenario as for the highest SAR values—55 V/m and 49 V/m, calculated according to ICNIRP and IEEE requirements, respectively, as shown in
Table 4. Additionally, in the case of the Ein analysis, no significant differences (up to 15% for various HF RFID reader locations) between the values in the model of a person using a PICC device obtained for all investigated exposure scenarios were found—all differences were below the level of uncertainty of the numerical simulations carried out, estimated as ±25% (K = 1). The 99th percentile value of Ein in all human body model tissues were up to 11 times lower than peak Ein values. Additionally, the Ein values calculated according to ICNIRP requirements were up to 20% higher than the values calculated according to IEEE requirements.
In the case of the bystander, the highest Ein values were found in the same exposure scenario as for a person using a PICC device—2.0 V/m and 1.5 V/m, calculated according to ICNIRP and IEEE requirements, respectively, as shown in
Table 4. These values were up to 45 times lower than the Ein values mentioned above, obtained in a model of a person using a PICC device, and up to 40% higher than the Ein values obtained for an HF RFID reader attached to the metal vehicle cabin. The 99th percentile values of Ein in all human body model tissues were up to 7 times lower than peak Ein values. Additionally, the Ein values calculated according to ICNIRP requirements were up to 40% higher than the values calculated according to IEEE requirements.
The highest Ein values (shown in
Table 4) in the model of a person using a PICC device reached up to 5% of the limits of GP and up to 2% of the limits off OE. All Ein values in the model of the bystander were below 1% of these limits.
4. Discussion
Various HF RFID applications in IoT systems are used with different RRs (typically from a few up to several centimetres, depending on particular application needs and the individual settings of output power in readers), as well as the use of PICCs of various classes (from 1 to 6, according to ISO/IEC 14443-2:2020).
Table 5 and
Table 6 present the results of the numerical simulations of SAR and Ein values, respectively, related to the exposure of a person using a PICC device to EMF at 13.56 MHz near to the HF RFID reader with an internal antenna with inner dimensions of 8 × 8 cm, assuming that the RR of the considered reader is 4, 10 or 16 cm and used various PICCs devices (of classes 1–6). The maximum RRs for which SAR or Ein values are compliant with limits of general public exposure (ICNIRP) or limits of exposure in unrestricted environments (IEEE) are also shown.
The SAR values calculated for systems with RRs of 10 and 16 cm were 30 times and 330 times higher than the SAR values calculated for an RR of 4 cm (
Table 5). Additionally, SAR values up to 2, 3 and 9 times higher were found for PICCs of classes 4, 5 and 6, respectively, compared to the values corresponding to the use of PICCs of classes 1–3.
It was found that only the model of a person using a PICC device of classes 5 and 6, in the system with 16 cm RR, experienced WBSAR and local SAR10g in head and torso values reaching up to 12% and 37%, respectively, of the limits of GP (shown in
Table 5). For those cases, the WBSAR and local SAR10g in head and torso values did not exceed 10% of the limits of OE. Additionally, local and SAR10g in limbs exceeding 115% of the limits of GP was only found in a person using a PICC device of class 6 in the system with 16 cm RR (23% of the limits of OE). For other cases in the system with 16 cm RR, those values reached 13%; 23% and 35% of the limits of GP for PICC devices of classes 1–3; 4 and 5, respectively. The values obtained for other investigated cases did not exceed 10% of the limits of GP and OE.
The maximum RRs for which SAR values were compliant with limits of GP exposure were estimated to be between approximately 23 cm (when the most common PICCs of classes 1–3 are used) and 15 cm (when PICCs of class 6 are used). The results of this study showed that SAR values in the body exposed to EMF at 10 cm away from HF RFID readers (in exposure scenarios similar to those considered in this study) may exceed the limits of GP exposure, when the RR exceeds 15 cm (taking into account continuous exposure over a time when anyone stays for longer than 6 min near the validator of continuous RF-EMF emission). Such a case of exposure is especially possible during rush hour and does not have to be related to the exposure while validating a public transport card. In cases of an exposure duration shorter than 6 min, the SAR values will be proportionally lower. However, it should be noted that, in an extreme case, the human body may even touch an HF RFID reader (a case of exposure not covered by this study) for which SAR values (especially local ones) may be significantly higher than those presented in this study. Additional numerical simulations are required to take that extreme case of exposure into consideration.
Moreover, the magnetic field strengths in the location of a person using a PICC device (undisturbed by the presence of a human) do exceed the reference levels limits of GP exposure in most of the considered exposure scenarios (and in many cases, the limits of OE are exceeded), but the SAR values are compliant with the limits, except for the exposure case when PICCs of class 6 are used in the system with an RR exceeding 15 cm.
The Ein values calculated for systems with RRs of 10 and 16 cm were up to 6 times and 18 times higher than Ein values calculated for an RR of 4 cm (
Table 6). Additionally, up to 30%, 70% and 200% higher Ein values were found for the use of PICCs of classes 4, 5 and 6, respectively, compared to the values corresponding to the use of PICCs of classes 1–3.
The Ein values (shown in
Table 6) in the model of a person using a PICC device of classes 1–6 in systems with 4 and 10 cm RRs did not exceed 10% of the limits of GP and OE, except for the PICC device of class 6 and 10 cm RR, for which they reached up to 15% of the limits of GP. In the case of a system with a 16 cm RR, Ein values reached up to 15%, 20%, 25% and 45% of the limits of GP and up to 7%, 9%, 12% and 20% of the limits of OE for the use of a PICC device of classes 1–3, 4, 5 and 6, respectively.
The maximum RRs for which Ein values are compliant with the limits of general public exposure (ICNIRP) or limits of exposure in unrestricted environments (IEEE) was estimated:
Between approximately 32 cm (most common PICCs of classes 1–3) and 22 cm (PICCs of class 6)—when considering ICNIRP 2020, peak values and limits for 10 MHz;
Between approximately 36 cm (PICCs of classes 1–3) and 25 cm (PICCs of class 6)—when considering ICNIRP 2020, peak values and limits extrapolated linearly for 13.56 MHz;
Between approximately 72 cm (PICCs of classes 1–3) and 50 cm (PICCs of class 6)—when considering ICNIRP 2010, 99th percentile values and limits for 10 MHz;
Between approximately 80 cm (PICCs of classes 1–3) and 55 cm (PICCs of class 6)—when considering ICNIRP 2010, 99th percentile values and limits extrapolated linearly for 13.56 MHz;
Between approximately 46 cm (most common PICCs of classes 1–3) and 31 cm (PICCs of class 6)—when considering IEEE, peak values and limits for 5 MHz;
Between approximately 63 cm (most common PICCs of classes 1–3) and 42 cm (PICCs of class 6)—when considering IEEE, peak values and limits extrapolated linearly for 13.56 MHz.
The results of this study showed that Ein values in a body exposed to EMF at a distance of 10 cm away from HF RFID readers in exposure scenarios similar to those considered in this study do not exceed the limits of GP exposure for typical reading ranges (up to 15–18 cm). Additionally, in the case of Ein assessment, extreme cases of exposure when the human body touches the HF RFID reader requires further investigation.
In public transport, any person present in the “ticket area” may be exposed to RF-EMF from three HF RFID readers, as shown at
Figure 1a. Such a case (called multiple-source exposure), along with cases of multifrequency exposure according to international guidelines and standards (ICNIRP, IEEE and IEC 62232:2017), needs to be included in the SAR and Ein assessment [
8,
10,
13,
15]. Guidelines on how to assess this exposure, depending on whether it is correlated or uncorrelated in time, are provided by IEC Technical Report 62630:2010 or ISO/IEC 14443-2:2020 [
15,
18], for example. According to these guidelines, under exposure from multiple sources uncorrelated in time, such as in the “ticket area”, SAR is an algebraic sum at any point in the body (mass element) of SAR values calculated separately for each considered source. At the worst theoretic case, assuming that all RF-EMF sources are in adjacent, very close locations, SAR and Ein values will be many times higher, as there are sources taken into account in the assessment, i.e., in the case of a person present in the “ticket area”—3 times higher (three HF RFID readers) than the values presented in this study. The maximum RRs for which SAR and Ein are compliant with the limits of GP, exposure will be approximately 15% and 30% shorter, respectively, than corresponding maximum RRs for a single HF RFID reader in this study (as shown in
Table 5 and
Table 6). However, under real (typical) exposure conditions of a person present in the “ticket area”, the SAR and Ein values in real environmental situations are expected to be similar to those presented above for a single reader of continuous EMF emissions (i.e., the location of readers by various parts of the body, greater distances among readers than their dimensions, continuous EMF emissions from readers built into the validator and short EMF emissions (from several seconds to tens of seconds) from readers built into ticket machines).
ISO/IEC 14443-2:2020 [
7] distinguishes two communication signal interfaces: types A and B. According to this standard, communications from HF RFID readers to PICC devices requires Amplitude Shift Keying (ASK) in 100% or 10% modulations. In the case of communications from PICC devices to HF RFID readers, other modulations are used, e.g., Binary Phase Shift Keying (BPSK) or On/Off Keying (OOK) as well as various coding such as Non-Return to Zero (MRZ) or Manchester, respectively. Furthermore, in communications from the reader to the PICC device, an ASK modulation of 2–3 μs duration should be used. The ratio between modulated and unmodulated periods of transmission time depends on the ratio of “1” and “0” in data to be transmitted. The duration of the transmission depends on data size and transmission rate. Another important parameter is the duty cycle. This depends on the purpose of the HF RFID system application and ranges from 0.1 to 0.8 according to the Electronic Communications Committee (ECC) report 208 [
19]. All these parameters influence the value of the emitted power averaged over time and thus SAR values (time averaged). The parameters influencing the SAR values the most are ASK 100% modulation and the duty cycles mentioned above. The SAR values should be 10 times, 2 times and 20% lower than the values presented in this study for continuous RF-EMF emissions when duty cycles of 0.1, 0.5 and 0.8 are considered, respectively. The maximum RRs for which SAR is compliant with the limits of GP exposure will be approximately 50%, 15% and 5% longer for duty cycles of 0.1, 0.5 and 0.8, respectively, than corresponding to maximum RRs for the HF RFID reader in this study.
However, it must be pointed out that the maximum RRs for which the considered Ein is compliant with the limits of GP exposure remain at the same level, because Ein values are not time averaged. Additionally, it was found that for low-duty cycles (e.g., 0.1), the maximum RRs for which Ein is compliant with the limits of GP exposure are shorter than RRs for which SAR is compliant with these limits. This justifies the need to evaluate not only SAR limits compliance but also assess the Ein compliance with limits under RF-EMF exposure from HF RFID readers. A detailed analysis of the impact of the communication parameters, the modulations, duty cycles, etc., also requires further investigation.
If HF RFID readers equipped with antennas with dimensions smaller than the dimensions used in this study are used in IoT systems, both the SAR and Ein values (both in the person validating the public transport card 10 cm away from HF RFID reader and the bystander) will be lower than those presented in this article, though anyone in direct proximity to the emitting antennas may be exposed to higher values.
To take another point of view, it is worth mentioning the work currently underway related to the application of IoT systems using HF RFID technology that does not require the PICCs to be brought near the readers (payment for journeys will be charged, for example, on the basis of identifying a proximity payment card, e.g., in a pocket of a passenger passing nearby the reader). Such a solution would require a significant increase in RRs, and thus an increase in exposure to the EMF emitted by them, along with an increase in the SAR and Ein values in people near to the readers.
It should be noted that people using public transport may also be users of active implantable medical devices (AIMDs) such as hearing implants, cardiac pacemakers, implantable cardioverter-defibrillators, insulin pumps, etc. Following requirements regarding the protection of the public and workers against electromagnetic hazards, potential malfunctions of AIMDs exposed to EMF from HF RFID readers need to be taken into consideration [
20,
21]. For example, published studies show higher SAR values in tissues next to AIMDs (their metal elements) compared to the SAR values calculated in the same tissues of a person who is not a user of such implants [
2,
22,
23,
24,
25]. Such an effect decreases the maximum RRs that may be counted to ensure that the general public limits of SAR and Ein are not exceeded in anyone present near HF RFID readers. Taken together, despite the results of my studies, which suggest that today, the EMF exposure in the ticket area of public busses seems to be typically compliant with relevant limits of general public exposure, further development in the applications of this technology also needs attention to the evaluation of the effects of EMF exposure, as well as attention to the possible technical measures reducing the EMF influence on anyone present nearby readers (especially when the use of readers of higher EMF emission is considered, even when today, the frequency of 13 MHz is not covered by the regular EMC requirements applicable for AIMD).
5. Conclusions
The main contribution of this work is that it shows that absorption in the human body of RF-EMF emitted continuously by a single HF RFID reader working in the IoT system in public transport vehicles may have a significant influence on humans when it is used in the system with RRs longer than approximately 15 cm, with PICCs devices of class 6 or longer than approximately 23 cm, with the most common PICCs devices of classes 1–3. It has also been shown that, in the bus ticket area, where having multiple sources of exposure is common (from three RFID readers installed there), the RF-EMF absorption in the human body may have a significant influence on anyone present there when RRs are 15% shorter than in the case of a single reader.
It should also be pointed out that, under RF-EMF exposure from HF RFID readers operating at any frequency from the HF band (3–30 MHz frequency), in exposure scenarios similar to those considered in this study (the body of person using a PICC device exposed to RF-EMF with the torso and hand holding the PICC device 10 cm and 0.5 cm away from the HF RFID, respectively, and a bystander present near to the HF-RFID-emitting antennas), the compliance assessment with public exposure limits of both SAR and Ein should be considered for low-duty cycles of RF-EMF emissions (at a level of 0.1).
This study indicates that the presence of metal objects, such as vehicle cabin, as well as people present in the vicinity of HF RFID readers can have an influence on the distribution of EMF emitted by the reader.
In this work, the main contribution concerned the HF RFID readers used in public transport vehicles, but similar devices are widely used in many other businesses, such as offices, shops, factories, medical centres and so on, and RF-EMF exposure and electromagnetic hazards in their vicinity are expected to be similar to that presented in this study.