Preprint, In: Proceedings of the 2026 IEEE 103rd Vehicular Technology Conference (VTC2026-Spring), Nice, France
Experimental Evaluation of LPWAN Technologies: mioty, LoRaWAN, Sigfox, NB-IoT, and LTE-M in Deep Indoor Environments Christof Röhrig, Benz Cramer
arXiv:2605.23483v1 [cs.NI] 22 May 2026
IDiAL – Institute for the Digital Transformation of Application and Living Domains Dortmund University of Applied Sciences and Arts Dortmund, Germany [email protected], 0000-0002-3286-3703
NB-IoT, LTE-M) operated by network operators with each other.
Abstract—Low Power Wide Area Networks (LPWAN) are often used in applications such as Smart City, Smart Buildings and Smart Metering. Energy meters are often located in underground spaces that are difficult to reach with wireless technology. This paper presents an experimental study comparing different LPWAN technologies in terms of building penetration. The technologies mioty® , Low Power Long Range Wide Area Network (LoRaWAN® ), Sigfox, Narrow Band Internet of Things (NB-IoT), and Long Term Evolution for Machines (LTE-M) are evaluated experimentally. The aim of the research is to investigate the performance of building penetration of different radio technologies in real-life scenarios. The measurements are performed with inexpensive off-the-shelf equipment. The results of the study are integrated in the metering system of the Dortmund University of Applied Sciences and Arts. Index Terms—LPWAN, smart metering, smart buildings
II. Problem definition and methodical approach This research compares five LPWAN radio technologies for smart metering in deep indoor and underground environments. The aim of the research is to investigate the performance of building penetration of different radio technologies in real-life scenarios. The measurements are performed with inexpensive off-the-shelf equipment. We focus on measuring the building penetration loss (BPL) without considering the power consumption of the equipment. The BPL LBP is a logarithmic attenuation measure that is specified in dB. We consider two scenarios: In the first scenario, the radio infrastructure, such as gateways (GWs) and base stations (BSs), is installed inside the building. Radio waves propagate through walls and ceilings. According to [2], the theoretic path-loss model for indoor environments has the form ! d Lindoor = L(d0 ) + N · log10 + L f (n), (1) d0
I. Introduction The Dortmund University of Applied Sciences and Arts operates several university buildings spread across the city of Dortmund. Energy meters are located in underground rooms and energy supply tunnels that are difficult to reach with wireless technology. The paper presents an experimental study that compares different LPWAN (Low Power Wide Area Network) technologies in terms of building penetration and radio coverage. These LPWAN technologies differ in terms of energy consumption, building penetration, radio coverage and radio range. The experiments will be carried out in various university buildings and in an underground parking garage with six levels half high. This paper extends the work we have presented in [1] and presents the following contribution: We evaluate five LPWAN technologies in demanding environments in terms of building penetration and radio coverage. The technologies mioty® , Low Power Long Range Wide Area Networks (LoRaWAN® ), Sigfox, Narrow Band Internet of Things (NB-IoT), and Long Term Evolution for Machines (LTE-M) are evaluated experimentally. In [1] we have compared LoRaWAN® with Wi-SUN® , Sigfox and NB-IoT. As mioty® becomes more widespread and offers greater reliability compared to LoRaWAN® , mioty® is being experimentally compared with LoRaWAN® . In a second environment we compare three LPWAN technologies (Sigfox,
where N is the distance power loss coefficient, d is the distance between the BS and the device, d0 is the reference distance, L(d0 ) is the path loss at d0 , L f is the floor penetration loss, n is the number of floors between BS and device (L f = 0 dB for n = 0). In the second scenario, the radio infrastructure (GWs, BSs) is installed in the outdoor environment. The radio waves propagate through the windows or the exterior wall into the building (Outdoor to Indoor, O2I). The penetration loss O2I depends heavily on the material, which the radio waves have to penetrate. The path loss incorporating O2I can be modeled LO2I = Lb + Ltw + Lindoor + N(0, σ2P ),
(2)
where Lb is the basic outdoor path loss, Ltw is the path loss through the external wall, Lindoor is the inside loss dependent on the depth into the building, and σP is the standard deviation for the penetration loss. Values for Ltw are given in [3], e.g. Lconcrete /dB = 5 + 4 · f /GHz and Lglass /dB = 2 + 0.2 · f /GHz.
1
Due to the frequency dependence of the BPL, low frequencies are preferred when good building penetration is required. An additional loss Lnpi must be added if the radio signal propagates through the external wall by non-perpendicular incidence [3]. It is very difficult to calculate the BPL based only on theoretical models because radio waves propagate in different ways in the building due to reflections and diffraction. In deep indoor and underground scenarios, such as basements and tunnels, the real signal path, and therefore the entire path loss, is too complex to be accurately represented by a linear model [4]. Therefore, we chose an empirical experimental evaluation where the received signal strength indicator (RSSI) is measured at outdoor measurement points (MPs) at ground level and then compared with the RSSI measured at indoor MPs on different floors and underground MPs in basements and (see Fig. 1). We measure the BPL
of smart urban drainage systems. Different depths of devices in manholes are investigated [16]. Thrane et al. present an experimental evaluation of an NB-IoT propagation model for deep indoor scenarios. The measurement campaign was conducted in a system of long underground tunnels [17]. IV. Technology Overview
LPWAN technologies known for good building penetration operate in the sub-GHz spectrum and include mioty® , LoRaWAN® , Sigfox, NB-IoT, and LTE-M. A large maximum coupling loss (MCL) is a prerequisite for good building penetration. The MCL describes the maximum difference between the transmitted and received power at the sender and receiver, respectively. mioty® , LoRaWAN® and Sigfox operate in the unlicensed spectrum of 868 MHz in Europe, while NB-IoT and LTEM operate in the licensed spectrum of 800 MHz (band 20) or 900 MHz (band 8) in Europe. To ensure fair use of the LBP = Poutdoor − Pindoor (3) unlicensed spectrum, each device is limited in transmit power as the difference between the received signal strength outdoors and duty cycle (DC). In the 868 MHz band, the transmit power Poutdoor and the received signal strength indoors Pindoor , both is normally limited to 14 dBm (25 mW) and a DC of 0.1%, measured in dBm. Aim of the study is to compare off-the-shelf except in the G1 band (868.0 - 868.6 MHz) where a DC of 1% is allowed and the G3 band (869.4 - 869.65 MHz) where devices in a real-world environment. a transmit power of 27 dBm (500 mW) and a DC of 10% is III. Related Work allowed. The G3 band is typically used for downlink (DL, LPWAN technologies have received increased attention in from BS to device) messages where a BS is connected to many recent years. They are characterized by their long range and low devices and transmits at high power. In contrast, quality of power consumption. Orlovs et al. provide a literature survey of service can be guaranteed in the licensed spectrum, where LPWAN technologies in real-world deployments [5]. LPWAN the mobile network operator (MNO) controls the use of the technologies known for their good building penetration include spectrum and the DC is not limited. mioty® is an LPWAN protocol that uses telegram splitting mioty® , LoRaWAN® , Sigfox, NB-IoT, LTE-M. ® mioty is a relatively young technology, which is why (TS) and ultra narrow-band (UNB) in the unlicensed spectrum. there are only a few experimental studies on mioty® in TS divides a data telegram into several sub-packets and real environments in the published literature. Robert and sends them after applying forward error correction (FEC) Lauterbach compare mioty® with LoRaWAN® with laboratory in a partially predefined time and frequency pattern. This measurements in [6]. Zerai et. al. present a comparative study makes the transmission robust against interferences and packet for LoRaWAN® and mioty® in an industrial environment [7]. collisions. Thanks to the FEC, up to 50% of sub-packets can be Joubert presents an experimental study comparing mioty® with compensated for. Each sub-packet requires a signal bandwidth LoRaWAN® at Schneider Electric in Grenoble [8]. Schreiber (BW) of 2.3 kHz and an air time of 15.14 ms in normal mode and Fosalau provide a theoretical comparision of mioty® with (EU1). TS is generated in a software-defined radio (SDR) NB-IoT with focus on metering applications [9]. Sikora et al. and uses standard Gaussian Minimum Shift Keying (GMSK) compare mioty® , LoRaWAN® , Sigfox and NB-IoT theoretically modulation, which can be produced using standard sub-GHz and provide an experimental evaluation in a unified testbed radio chips. The standard is defined in the specification TS 103 [10]. Oberacher et. al. report experiences from a large-scale 357 of the European Telecommunications Standards Institute (ETSI). In Europe, mioty® is using the G1 band for the uplink mioty® installation [11]. ® Kadusic et al. provide an overview of LoRaWAN , NB-IoT (UL, from device to BS) and the G3 band for the DL. It requires and Sigfox and the characteristics of these radio technologies a BW of 200 kHz for two channels in UL and DL (EU1). The [12]. Naumann and Oelers compare the energy consumption data rate of mioty® is 0.5 kbit/s, the payload of a telegram can and the link budget of NB-IoT, LoRaWAN® and Sigfox based be up to 250 bytes. The strengths of mioty® are its low power on theoretical models [13]. Trendov et. al. present a comparative consumption of 17.8 µWh (device, EU1) per message and its study for LoRaWAN® , NB-IoT, Sigfox and LTE-M in urban ability to support an ultra-high device density with up to 3.5 outdoor environments (cities of Halle and Köthen) [14]. Stusek million messages per day and BS [9]. Further details on the et al. provides a comparative study of LPWAN propagation physical layer of mioty® can be found in [6]. models in urban scenarios [15]. They compare propagation LoRaWAN® is based on the proprietary LoRa® radio tech® models for LoRaWAN , Sigfox, and NB-IoT. Roosipuu et nology developed by Scemtech. LoRaWAN® defines the comal. investigate the use of NB-IoT for monitoring and control munication protocol and system architecture and is managed
2
by the open LoRa Alliance® . It can achieve data rates between 0.25 kbit/s and 11 kbit/s in Europe. It uses spreading factors (SF) to adjust the data rate and receiver sensitivity. A higher SF increases sensitivity, but reduces data rate. The payload of the messages can be from 51 to 222 bytes in Europe depending on the data rate. The number of messages is limited by the DC regulations in Europe and may be even lower due to restrictions by the LoRaWAN® network operator. The Things Network’s (TTN) fair use policy limits a device’s airtime to 30 s per day. LoRaWAN® networks can be operated by a network operator, a community such as TTN, or self-deployed. LoRaWAN® thus offers a high degree of flexibility for the end user. The proprietary Sigfox radio technology was developed by the French company Sigfox S.A. (now owned by UnaBiz). It uses UNB with a signal BW of 0.1 kHz and achieves long range and requires low power. The Sigfox network is based on a star topology and requires a local Sigfox network operator to carry the traffic generated. In Europe, Sigfox is using the G1 band for the UL and the G3 band for the DL. Each message is transmitted by default three times on different, randomly selected frequencies. This channel accessing scheme is named Random Frequency Division Multiple Access (RFDMA). The demodulation in the BS is done by SDR which analyzes the total band, to detect transmitted signals and to retrieve sent data. Sigfox supports up to 140 UL and 4 DL messages per day, each carrying a payload of 12 bytes and 8 bytes respectively, at a data rate of 0.1 kbit/s. Sigfox subscription cost is 10 € per year for one device and 140 UL, 4 DL messages per day (Heliot Europe). NB-IoT operates in licensed spectrum and is standardized by the 3rd Generation Partnership Project (3GPP) as LTE Cat-NB1 and -NB2. It is only available through MNOs. Implementing NB-IoT is cost effective as it is based on existing cellular infrastructure. Compared to other LPWAN technologies that operate in unlicensed spectrum, the transmit power and DC are not limited by regulation. Therefore, the MCL is higher, resulting in better building penetration. The downside is higher power consumption. NB-IoT offers data rates of up to 62 kbit/s in Cat NB1 and up to 159 kbit/s in Cat NB2 in the UL and up to 26 kbit/s and 127 kbit/s respectively in the DL. Applications are typically low throughput, delay tolerant and low mobility. Examples are smart meters and remote sensors. The cost of the subscription is 11 € for a SIM card, 500 MByte for 10 years (1NCE, IoT Lifetime Flatrate). LTE-M operates in licensed spectrum and is standardized by the 3GPP as LTE Cat-M1 (3GPP release 13) and Cat-M2 (3GPP release 14). Compared to NB-IoT the data rate is large in UL and DL: 1 Mbit/s for Cat-M1 and 4 Mbit/s - 7 Mbit/s for Cat-M2. The latency is much lower compared to NB-IoT: 10 – 15 ms for Cat-M1 and 50 – 100 ms Cat-M2. The downside is the lower receiver sensitivity and therefore a lower MCL. The costs of subscription are the same as for NB-IoT at the same data volume (1NCE). Table I compares the key parameters of the technologies (at highest sensitivity and MCL). The MCL is the maximum possible path loss from transmitter to receiver. A high value is
important for penetrating underground environments. The values are taken from [13] and the data sheets of the experimental equipment in Table II. TABLE I Comparison of different LPWAN technologies (Europe)
Technology Band MHz Tx pow. GW2 Tx pow. dev.2 Rx sens. GW2 Rx sens. dev.2 MCL UL3 MCL DL3 BW4 price module annual cost Dev. availab.
mioty® 868 27 14 -1355 -128 149 155 2.3 15€ 0
1: depending on the frequency of use,
2: in dBm,
LoRa® 868 14/271 14 -140 -136 154 150/1631 125 6€ ++ 3: in dB,
Sigfox 868 27 14 -142 -126 156 153 0.1 8€ 10€ -
Cat-NB / M 800 / 900 23 23 -141/-131 -139/-129 164/154 162/152 3.75/15 7€/14€ 1-2€ +
4: in kHz, 5: depending on the SDR frontend
The sensitivity of a receiver to a particular LPWAN technology is specified at a given signal-to-noise ratio (SNR). The SNR limit depends on the modulation scheme of the technology. The LoRa® receiver sensitivity is −140 dBm at the GW with an SNR of −20 dBm. The Sigfox receiver is sensitive to signals with an SNR of −142 dBm at the BS with an SNR of 9 dBm. The SNR values can not be compared directly because the BW of the technologies differs large. LoRaWAN® uses a BW B = 125 kHz and Sigfox B = 0.1 kHz. For thermal noise density N0 = −174 dBm/Hz, the thermal noise level is −123 dBm at B = 125 kHz and −154 dBm at B = 0.1 kHz. Therefore, at the same noise density, Sigfox’s noise power is 31 dB lower than LoRa’s. Due to the different BWs of the various LPWAN technologies, the SNR is not compared experimentally. V. Experimental evaluation of LPWAN technologies A. Experimental setup We use several off-the-shelf devices for the experiments (see Table II). This equipment does not provide an accurate measurement of RSSI, but is intended to provide an estimate of building penetration using off-the-shelf devices. The receiver sensitivities in Table II are taken from the data sheets and represent the best-case scenario for low noise. In a noisy environment, however, the sensitivity of the receiver can be much lower than that published in the data sheets. The mioty® AVA GW provides an estimate of the receiver sensititvity based on noise measurements. In our experimental environment, the sensitivity of the GW is estimated to be −125 dBm, based on a measured noise density of −156 dBm/Hz. The sensitivity undergoes a change over time due to its dependence on the fluctuating noise power. The experiments were conducted in two different environments. As there is poor Sigfox coverage on the university campus (see [1]), for the first measurements an environment was chosen in which Sigfox, NB-IoT and LTE-M could be received with a strong signal. For mioty® and LoRaWAN® , an infrastructure with dedicated GWs was set up on the campus.
3
TABLE II Devices used in experiments
Device type manufacturer radio chip technology band MHz Tx pow. / dBm Rx sen. / dBm
AVA GW Weptech SDR mioty® 868 14 -135
MUNIA-M device Weptech TI CC1310 mioty® 868 14 -128
M3B Magnolinq device LZE GmbH TI CC1310 mioty® 868 14 -128
LPS8 GW Dragino SX1308 LoRa 868 14 / 20 -140
B. O2I penetration of Sigfox, NB-IoT and LTE-M (public infrastructure)
LoRa E5 mini device Seeed STM32WLE LoRa 868 14 -136.5
MKR FOX device Arduino ATAB8520E Sigfox 868 13 -126
SIM7080G device Simcom MDM9205 NB-IoT / LTE-M 900 / 800 20 -130 / -107
TABLE III RSSI values in dBm and LBP in dB for Sigfox, NB-IoT and LTE-M level O A
The first measurement campaign was conducted in an underground car park with six half underground levels (A-F) in the centre of Dortmund. In this environment, the Sigfox, NBIoT and LTE-M BSs are close to the car park, resulting in high outdoor RSSI values. The distance between the experimental environment and the NB-IoT/LTE-M BS is approximately 100 m. The location of the Sigfox base station is unknown. Since the RSSI level received from the Sigfox BS is relatively high (outdoors: −61 dBm), the BS is likely not far away. The signal received from the next public LoRaWAN® GW (distance 800 m) is weak (outdoors: −96 dBm), therefore LoRaWAN® is not included in this experimantal study. We utilize commercial LPWAN networks. For Sigfox Heliot provides the network, for NB-IoT and LTE-M, Dt. Telekom is choosen. Measurements are made on different underground floors to measure the penetration loss of the ceilings. The signal strength values for NB-IoT and LTE-M are measured DL and are obtained directly from the modem using AT commands. The device SIM7080G provides several signal strength and quality parameters. In this research, the Reference Signal Received Power (RSRP) is choosen, because it provides more accurate signal power estimations by excluding interference from other sectors. The Sigfox RSSI values are measured by the BS in UL direction and reported by the Sigfox back-end system. Table III shows the results of the experimental evaluation. The RSSI values for level 0 are measured outdoors at the entry of the car park and are the reference values for calculating LBP using (3). In this measurement campaign, NB-IoT provides better building penetration than LTE-M and Sigfox, which cant’t reach the lowest level F. LBP is significantly higher for Sigfox compared to NB-IoT and LTE-M. This could be caused by a flatter angle of entry of the signal into the car park, which would result in a higher attenuation of the signal.
C E F
RSSI RSSI LBP RSSI LBP RSSI LBP RSSI LBP
Sigfox -61 -103 42 -113 52 -134 73 -
NB-IoT -53 -84 31 -95 42 -112 59 -123 70
LTE-M -60 -87 27 -98 38 -112 52 -
GW LPS8 and the mioty® GW AVA is placed on the first floor near a window (see Fig. 1). Two MPs are used as references: MP I is located near the GWs, and MP O is positioned outside, between the three buildings. The naming scheme for the other MPS is <building>.<level>.<point>, where ’B’ denotes the basement. Fig. 1 shows the experimental area at the campus Emil-Figge-Str. in Dortmund. The measured RSSI values at the outdoor MP O correspond to the basic outdoor path loss Lb in (2). At each MP, 20 measurements are made for mioty® . For LoRaWAN® , 27 measurements were recorded at each MP, 9
TABLE IV RSSI values in dBm, LBP in dB and PER in % for mioty® and LoRaWAN®
MP I 1.0.b 1.1.b 1.2.b 1.0.a 1.1.a 1.2.a 1.b O 2.0.a 2.B.a 2.0.b 3.0.b 3.B.b 3.B.a
C. Experimental evaluation of mioty® and LoRaWAN® in a private infrastructure In this measurement campaign, we compare mioty® with LoRaWAN® . Since there is currently no publicly available mioty® network, a private network was set up for this measurements. The experiments are performed on university campus Emil-Figge-Str. in Dortmund. In building 1 the LoRaWAN®
4
RSSI -42 -72 -93 -103 -86 -98 -112 -120 -66 -112 -99 -124 -123 -
mioty® PER 0 0 0 0 0 0 0 0 0 0 100 0 55 40 100
LBP 30 51 61 44 56 70 78 46 33 58 57 -
LoRaWAN® RSSI PER LBP -34 0 -71 0 37 -102 0 68 -108 0 74 -87 0 53 -103 0 69 -102 0 68 -129 33 95 -69 0 -120 0 51 -129 38 60 -95 0 26 -114 11 45 -123 11 54 -128 11 59
70
RSSI (in dBm)
80 90 100 110 120
2 mi
oty mi
oty
1
7 SF
8 SF
9 SF
10 SF
11 SF
SF
12
130
Fig. 2. Boxplot of measured RSSI values for mioty® ans LoRaWAN®
values are not comparable due to the fact that the mioty® devices utilise PCB antennas, whereas the LoRaWAN® device is equipped with a small rod antenna. In comparison with the LoRaWAN® device, the two mioty® devices exhibit a lower PER in relation to the mean RSSI value. In case of SF12 the LoRaWAN® GW provides a better sensitivity and therefore higher MCL compared to mioty® . The disadvantage of a high SF lies in the longer occupation of the frequency band and higher energy consumption. LoRaWAN® offers a method to automatically optimize data rate, airtime, and energy consumption with Adaptive Data Rate (ADR).
Fig. 1. Experimental area at campus Emil-Figge-Str. with MPs
each for SF10, SF11, and SF12. The results are shown in Table IV. The UL RSSI values for LoRaWAN® are obtained from the TTN back-end, the UL RSSI values for mioty® are obtained directly from the mioty® AVA GW. The RSSI values in the Table IV are the medians of the measurements at the indicated MP. The packect error rate (PER) is calculated for mioty® and LoRaWAN® for all measured values at each MP. LoRaWAN® can reach the basements in all buildings, mioty® can not reach two MPs in basements of building 2 and 3. The lowest RSSI value measured by the mioty® GW is −125 dBm, the lowest RSSI value observed by the LoRaWAN® GW is −138 dBm. To compare LBP of mioty® with LoRaWAN® , the RSSI of the MPs is subtracted from the RSSI of the reference MP. Table IV compares LBP for each indoor MP. Due to measurement inaccuracies, these values are not exact. Furthermore, RSSI values are determined using different methods, which is why it is not possible to rank the performance of the technologies in relation to BPL. However, the measured values can be used to determine the order of magnitude of BPL to be expected.
TABLE V RSSI values in dBm, SNR in dB and PER in % for mioty® and LoRaWAN® RSSI SF7 SF8 SF9 SF10 SF11 SF12 mioty1 mioty2
max -89.0 -90.0 -90.0 -71.0 -86.0 -64.0 -93.7 -106.5
median -97.0 -97.0 -101.0 -96.0 -95.0 -97.0 -113.1 -118.9
min -122.0 -124.0 -126.0 -127.0 -130.0 -132.0 -128.3 -129.2
@SNR -9.5 -11.8 -15.2 -14.0 -16.8 -17.2 -2.5 -4.0
PER 12.6 8.8 5.1 4.0 4.3 3.1 1.4 10.7
Fig 2 shows a boxplot of the measure RSSI values. The resolution of RSSI values is 1 dBm for LoRaWAN® , while it is 0.1 dBm for mioty® . LoRaWAN’s RSSI values fluctuate more significantly than those of mioty® , particularly for SF12.
D. Long-term evaluation of mioty® and LoRaWAN® In order to evaluate the long-term fluctuation of RSSI values and the PER, two MUNIA-M devices and one LoRaWAN® device are deployed in building 1 of the experimental environment (see Fig 1). The device mioty1 has been installed in a laboratory situated on the third floor. The device mioty2 and the LoRaWAN® device have been installed in the server room that is located in the basement. The GWs have remained in the same location as in the previous experimental evaluation. The mioty® devices send every 15 minutes a message, the LoRaWAN® device transmits packets every 15 minutes, each with SF7 through SF12. In the course of a month, more than 2,600 messages were sent from each device. The results of the experiment are displayed in Table V. The absolute RSSI
VI. Conclusions The paper presents an experimental evaluation of five LPWAN technologies in deep indoor environments. NB-IoT, mioty® and LoRaWAN® show better building penetration than Sigfox and LTE-M. Because the density of Sigfox BSs is relatively low, indoor coverage depends heavily on the distance to the nearest BSs. NB-IoT offers higher reliability than other LPWAN technologies by using licensed radio bands and extended coverage levels (ECLs) with repetitions. Compared to NB-IoT, mioty® and LoRaWAN® have the advantage of
5
lower power consumption. LoRaWAN® offers the widest range of energy monitoring devices and the greatest infrastructure flexibility compared to other LPWAN technologies. mioty® is more reliable than LoRaWAN® if the RSSI values are high enough. In our experiments, the sensitivity of the mioty® GW is lower than that of the LoRaWAN® GW. This confirms the specifications listed in the data sheets for the GWs used. mioty® packet loss occured when the average RSSI value drops below −120 dBm. Since the sensitivity of a mioty® GW depend mainly on the SDR frontend, this may be a limitation of the mioty® GW under test. mioty® is a relatively young technology that is developing rapidly. Currently, there are fewer providers offering components for mioty® compared to LoRaWAN® , but mioty® may become more important in the future due to its higher reliability and ability to support a higher device density compared to LoRaWAN® . The results of the study are integrated in the metering system of the Dortmund University of Applied Sciences and Arts (see [18]).
[9] D. Schreiber and C. Fosalau, “LPWAN protocol selection for IoT-based energy measurement systems: A comparative analysis of mioty and NB-IoT technologies,” in 2025 International Conference on Electromechanical and Energy Systems (SIELMEN), 2025, pp. 046–051. [10] J. S. E, A. Sikora, M. Schappacher, and Z. Amjad, “Test and Measurement of LPWAN and Cellular IoT Networks in a Unified Testbed,” in 2019 IEEE 17th International Conference on Industrial Informatics (INDIN), vol. 1, Jul. 2019, pp. 1521–1527. [11] M. Oberascher, C. Maussner, P. Hinteregger, J. Knapp, A. Halm, M. Kaiser, W. Gruber, D. Truppe, E. Eggeling, and R. Sitzenfrei, “Using digitalisation for a real-word implementation of an early warning system for water leakages,” Water Practice and Technology, vol. 20, no. 1, pp. 49–64, 2024. [12] E. Kadusic, C. Ruland, N. Hadzajlic, and N. Zivic, “The factors for choosing among NB-IoT, LoRaWAN, and Sigfox radio communication technologies for IoT networking,” in 2022 International Conference on Connected Systems & Intelligence (CSI), 2022, pp. 1–5. [13] H. Naumann and W. Oelers, “LPWAN comparison – low energy consumption with NB-IoT, LoRaWAN and Sigfox,” 2021. [14] S. Trendov, E. Sariiev, K. B. S. Mukhtar, D. Kachan, and E. Siemens, “LPWAN Technologies Range Evaluation: LoRaWAN, NB-IoT, Sigfox, and LTE-M in Urban and Free-field Settings,” in 2025 7th Global Power, Energy and Communication Conference (GPECOM), 2025, pp. 1089–1094. [15] M. Stusek, D. Moltchanov, P. Masek, K. Mikhaylov, O. Zeman, M. Roubicek, Y. Koucheryavy, and J. Hosek, “Accuracy Assessment and Cross-Validation of LPWAN Propagation Models in Urban Scenarios,” IEEE Access, vol. 8, pp. 154 625–154 636, 2020. [16] P. Roosipuu, I. Annus, A. Kuusik, N. Kändler, and M. M. Alam, “Monitoring and control of smart urban drainage systems using NB-IoT cellular sensor networks,” Water Science and Technology, vol. 88, no. 2, pp. 339–354, 07 2023. [17] J. Thrane, K. M. Malarski, H. L. Christiansen, and S. Ruepp, “Experimental evaluation of empirical NBIoT propagation modelling in a deep-indoor scenario,” in GLOBECOM 2020 - 2020 IEEE Global Communications Conference, 2020, pp. 1–6. [18] C. Röhrig, D. Heß, and B. H. D. Trinh, “System design for distributed energy management using multiple LPWAN technologies,” in IECON 2025 - 51th Annual Conference of the IEEE Industrial Electronics Society, Madrid, Oct. 2025, pp. 1–7.
References [1] C. Röhrig and B. Cramer, “LPWAN based IoT architecture for distributed energy monitoring in deep indoor environments,” in Proceedings of the 6th International Conference on Building Energy and Environment (COBEE 2025), Eindhoven, Netherlands, Jul. 2025, pp. 1404–1411. [2] ITU-R, “Propagation data and prediction methods for the planning of indoor radiocommunication systems and radio local area networks in the frequency range from 300 MHz to 450 GHz,” P Series Radiowave propagation, Recommendation ITU-R, no. 1238–13, 2025. [3] 3GPP, “Study on channel model for frequencies from 0.5 to 100 GHz,” 3GPP TR 38.901 version 19.1.0, no. (Release 19), 2025. [4] K. M. Malarski, J. Thrane, M. G. Bech, K. Macheta, H. L. Christiansen, M. N. Petersen, and S. Ruepp, “Investigation of deep indoor NB-IoT propagation attenuation,” in 2019 IEEE 90th Vehicular Technology Conference (VTC2019Fall), 2019, pp. 1–5. [5] D. Orlovs, A. Rusins, V. Skrastin, š, and J. Judvaitis, “LPWAN Technologies for IoT: Real-World Deployment Performance and Practical Comparison,” IoT, vol. 6, no. 4, 2025. [6] J. Robert and T. Lauterbach, “LoRa and Mioty - A Comparative Study,” in Global Internet of Things and Edge Computing Summit, M. Presser, A. Skarmeta, and S. Krco, Eds. Springer Nature Switzerland, 2026, pp. 18–38. [7] O. Zerai, M. Striegel, and T. Krone, “LoRaWAN and MIOTY: A Study on Packet Reception and Energy Consumption in the Industrial Internet of Things,” ifm electronic GmbH, Tech. Rep., Jun. 2023. [8] A. Joubert, “Mioty vs LoRaWAN – On Site Field Test,” https://hub.imt-atlantique.fr/lpwan25/presentations/ 4/session1/4.pdf, LPWAN Days 2025.
6