2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)

Study of LoRa Performance at 433 MHz and 868

MHz Bands Inside a Multistory Building
Ivan Bobkov Alexey Rolich
National Research University Higher School of Economics National Research University Higher School of Economics
Moscow, Russia Moscow, Russia
ivbobkov@miem.hse.ru arolich@hse.ru
Maria Denisova Leonid Voskov
National Research University Higher School of Economics National Research University Higher School of Economics
Moscow, Russia Moscow, Russia
madenisova_1@miem.hse.ru lvoskov@hse.ru

Abstract—Owing to its energy efficiency and long range, This paper describes an architecture of a LoRa indoor
LoRa wireless network has become a widely spread technology network experimental stand for carrying out experiments
among IoT systems recently. LoRa allows to use various ISM with the transfer of packets indoors under various
bands such as 433 MHz, 868 MHz and 915 MHz. During this transmission conditions and the results of experimental
study 433 MHz and 868 MHz frequencies have been compared.
studies of the received signal level indicator for 433 MHz
Parameters such as SNR and RSSI were measured at different
floors and visualized. A comparative table of packet delivery and 868 MHz at various spreading factors.
ratio at various spreading factors can be found in this paper. A Results are presented in the following order: Section II
series of range experiments at different spreading factors contains an overview of related works. The rest of the paper
showed that 433 MHz LoRa module gains a stronger signal.
However, 868 MHz LoRa module shows higher percentage of
is organized as follows: section III briefly describes LoRa
received packets. It has been concluded that for nine-story network experimental stand architecture. Furthermore, the
building with concrete floors it is better to deploy 868 MHz conducted experiments are presented and described in
LoRa network at 10th spreading factor. section IV. Section V contains the received data and its
analysis. Finally, the obtained results of the study are
Keywords—IoT, Internet of things, LoRa, indoor concluded in section VI.
performance, signal propagation, ISM band comparison, SNR,
RSSI, PDR, packet loss, spreading factor.
II. LITERATURE REVIEW
Islam, B. et al. [2] analyzed the feasibility of using
I. INTRODUCTION LoRa, an emerging low-power wide-area networking
Due to Internet of Things (IoT) constant development, technology, in indoor localization. Considering the
wireless technologies are actively infiltrating into our coverage, stability and regularity of signals, accuracy of
everyday life. As well as personal IoT devices, complicated localization, responsiveness, power, and cost – authors
IoT systems such as smart buildings are becoming more concluded that LoRa is a feasible choice for indoor
spread nowadays. Increasing smart buildings demand localization, especially in wide and tall indoor environments
requires innovative and more energy efficient solutions for like warehouses and multi-storied buildings.
information transmission. One of the most common network
In [5], authors present the indoor performance of LoRa
solutions for IoT systems is LoRa wireless communication
by the means of real-life measurements. The measurements
technology [1]. Compared with Wi-Fi and Bluetooth Low
were conducted using the commercially available equipment
Energy (BLE), LoRa is considered as a better choice for
in the main campus of the University of Oulu, Finland. The
indoor deployment in multi-story buildings [2].
paper [6] consider data collection in the Hanover building at
The subject of LoRa indoor placement is an important Glasgow Caledonian University using the LoRa
issue for IoT in smart buildings. There are several studies transceivers. This paper facilitates the link budget design,
where various LoRa indoor experiments were conducted network implementation and coverage diagnosis in similar
such as a research of LoRa technologies possibility in the indoor scenarios.
indoor environment in the context of human-centric
Erbati, Schieli and Batke [10] present behavior analyzes
applications [3–4] or a study of LoRa deployment
of LoRaWAN technology in urban settings in both an
specificity inside a flower auction warehouse [5]. Mainly,
indoor and an outdoor scenario, with respect to parameters
LoRa indoor articles concentrate on the measurement of
such as RSSI, SNR, PDR, Coverage, and Deep Indoor
received signal strength indicator (RSSI) [6–7], packet
Penetration (DIP).
delivery ratio (PDR) [8–9] and on the comparison of
different spreading factors [10–11]. However, there are only LoRa modules mainly use non-licensed 433 MHz, 868
a few papers where these significant parameters are MHz and 915 MHz industrial, scientific and medical (ISM)
compared at different frequencies. radio bands [11]. In this study 433 MHz and 868 MHz

978-1-7281-2572-5/20/$31.00 ©2020 IEEE

2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)

frequencies are compared indoors. To our knowledge, there

is only one research where these two ISM bands have
opposed each other inside a modern multi-story
construction. Based on the signal-to-noise ratio (SNR)
measurements through two thick concrete cores the article
claims that 434 MHz frequency is more preferable for LoRa
indoor communication [13] which this study partly proves.
Each standard LoRa module includes a spreading factor
(SF) – a customizable parameter which can be changed from
Fig. 2: Heltec Wi-Fi LoRa 32 (v2) module
7 to 12 conventional units. Recent study shows that the
greater the distance outdoors from transmitter to receiver is, Transmission nodes also include one battery holder for 4
the greater SF value should be. The article [14] compares AA batteries and are placed inside a plastic case with
different spreading factors indoors and illustrates the best antenna on the outside.
solution for multi-story buildings, depending on the number
of floors between LoRa modules.
In summary, all of the aforementioned works study the
feasibility of LoRa for indoor communication, formulating
many open issues and addressing some of them. Thus a
potential gap in study of LoRa indoor communication using
433 MHz and 868 MHz with different spreading factors can
be observed.

III. EXPERIMENTAL STAND ARCHITECTURE

A complete architecture of the provided experimental Fig. 3: LoRa transmission node
stand is presented in Figure 1. Stand includes the following
elements: During each experiment LoRa transmitter is
continuously sending packets which contain their
1. End-point data collection devices (nodes) using 433 identification number, spreading factor and number of floors
or 868 MHz frequencies (one or several sensors can between the transmitter and the base station. In total,
be connected to these devices; data transmission is transmitter sends 3000 packets per experiment (500 packets
performed according to the LoRa specification). for each of 6 SF). Switching between spreading factors
2. LoRa base station (consists of Heltec LoRa module happens automatically.
and Raspberry Pi 3). Packet reception takes place at LoRa base station which
consists of Heltec LoRa module described in subsection A
(v1 or v2 depending on the experiment frequency) and a
single-board computer Raspberry Pi 3 [16] which sends
collected (packet id, SF, floor difference) and calculated
(SNR, RSSI) data to server. LoRa receiver listens to the
broadcast from 7 to 12 SF consequentially.

Fig. 4: LoRa base station

Fig. 1: LoRa network experimental stand architecture IV. CONDUCTED EXPERIMENTS

Each LoRa node for signal transmission consists of a A. Nodes placement
LoRa module based on ESP32 microcontroller and SX1276
Experiments took place in a nine-story building of HSE
transceiver. These modules are provided by Heltec
Electronics and Mathematics Faculty. LoRa base station
Automation [15]: Wi-Fi LoRa 32 module for 433 MHz and
was permanently placed on the ground floor, while LoRa
Wi-Fi LoRa 32 (v2) module for 868 MHz experiments
transmitter was being moved from the second to the ninth
respectively.
floor consequentially. The transmitter was placed on the
concrete floor above the receiver as vertically as possible to
minimize deviations from a straight path.
2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)

B. Experiments description 4
-114.80 -115.72 -116.17 -117.80 -118.84 -118.41
The goal of the study was to find out which one of the
5
two frequencies is more suitable for LoRa indoor -116.73 -117.76 -118.26 -119.34 -119.80 -119.69
communication: 433 MHz or 868 MHz. Firstly, LoRa 433 6
MHz transmitter sent 3000 packets (500 packets for each -115.89 -116.50 -117.07 -118.44 -119.18 -119.36

SF) from every floor to the base station. Secondly, LoRa 7

-117.71 -118.50 -119.01 -120.77 -120.63 -120.63
868 MHz transmitter has done the same. Calculated SNR, 8
RSSI and packet loss have been analyzed for every SF of -120.35 -121.19 -121.54 -123.10 -122.21 -119.51
each ISM bands and presented as comparative diagrams in
section V. TABLE III. AVERAGE SNR FOR 868 MHZ
SF
7 8 9 10 11 12
Floor
1
9.55 9.05 9.20 9.16 9.15 9.15
2
8.88 8.97 8.96 8.96 8.97 8.98
3
8.92 8.95 9.00 8.96 8.99 9.00
4
9.01 9.08 9.10 9.05 9.04 8.95
5
9.11 9.13 9.17 9.21 9.22 9.21
6
9.02 9.06 9.11 9.10 9.08 9.16
7
9.00 9.07 9.08 9.08 9.06 9.10
8
9.07 9.09 9.13 9.05 9.09 9.13

TABLE IV. AVERAGE RSSI FOR 868 MHZ

Fig. 5: Nodes placement
SF
V. RESULTS AND ANALYSIS 7 8 9 10 11 12
Floor
A. 433 MHz and 868 MHz comparison
1
-77.83 -81.12 -80.18 -85.30 -86.51 -86.20
TABLE I. AVERAGE SNR FOR 433 MHZ
2
-109.54 -112.03 -108.99 -109.47 -110.91 -110.78
SF
7 8 9 10 11 12 3
-117.66 -118.22 -120.22 -120.07 -120.96 -121.50
Floor
4
1 -127.27 -127.50 -127.67 -129.22 -129.86 -129.77
12.47 12.40 12.44 12.43 12.56 12.47 5
2 -128.98 -129.45 -130.07 -130.91 -131.11 -130.91
12.30 12.39 12.40 12.49 12.38 12.39 6
3 -128.89 -129.54 -129.76 -130.89 -131.29 -131.10
12.28 12.37 12.33 12.35 12.35 12.33 7
4 -127.63 -128.13 -128.59 -130.26 -130.66 -130.63
12.30 12.36 12.37 12.36 12.38 12.38 8
-128.91 -129.29 -129.78 -131.08 -131.23 -131.13
5
12.80 12.66 12.70 12.79 12.79 12.82
6
12.67 12.72 12.74 12.73 12.73 12.71 TABLE VI. AVERAGE PDR FOR 433 MHZ
7
12.64 12.69 12.69 12.68 12.72 12.69 SF
8 7 8 9 10 11 12
12.27 12.44 12.45 12.46 12.57 12.85 Floor

TABLE II. AVERAGE RSSI FOR 433 MHZ 1

99.8 99.8 100 100 87 99.8
SF 2
100 100 99.8 100 100 100
7 8 9 10 11 12 3
Floor 77.2 75 94 99.2 99.6 100
1 4
-75.39 -76.50 -77.36 -80.96 -82.21 -80.48 99.4 99.6 100 100 99.4 100
2 5
-96.77 -97.35 -97.39 -99.53 -101.15 -101.45 37.2 97 98 98.8 98.4 99.2
3 6
-117.23 -118.91 -119.20 -120.04 -119.89 -119.90 99.8 99.8 99.4 100 100 98.8
2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)

7
82 92.2 95 93.2 99.8 100
8
32.2 79 88.2 98 99.2 99.2

Arithmetical mean values of all gathered SNR and

RSSI have been calculated for each floor and SF, then put
on the line graph. According to the provided comparison, it
can be concluded that LoRa shows better SNR and RSSI
values at 433 MHz than at 868 MHz ISM band.
TABLE VI. AVERAGE PDR FOR 868 MHZ
SF
7 8 9 10 11 12
Floor (b)
1 Fig. 6: Comparison of 433 MHz and 868 MHz SNR (a) and
99.8 99.8 99 100 99.8 100
RSSI (b)
2
99.2 100 93.6 100 100 99.8
3
100 99.8 97.6 100 100 100
4
98.6 99.8 99 100 99.8 99.6
5
79.4 87.4 92.8 99.6 99.4 99.4
6
97.4 99.4 100 99.4 99.6 99.8
7
98.8 100 100 100 100 99.4
8
97.60 99.6 99.8 100 100 99.4

However, comparison of packet delivery ratio (PDR)

illustrates more serious packet loss at 433 MHz frequency.
It can be observed that the lowest point of 433 MHz at the Fig. 7: Comparison of 433 MHz and 868 MHz packet
5th floor is much deeper than the bottom point of 868 MHz delivery ratio
ISM band (37.2% of packets were received at 7 SF against
79.4% respectively). The same situation occurs at the 8th B. Spreading factor analysis
floor (32.2% against 97.6%). Moreover, 433 MHz
percentage fluctuates more dramatically, while 868 MHz
percentage remains stable most part of the floors.

(a)

(a)
2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)

total, more stable transmission than 433 MHz ISM band (3

stable SF against 2, 1 unstable SF against 2).

VI. CONCLUSION
This paper proposed an architecture of a LoRa indoor
network experimental stand for carrying out experiments
with the transfer of packets indoors under various
transmission conditions and the results of experimental
studies of the received signal level indicator for 433 MHz
and 868 MHz at various spreading factors.
The study results show that 433 MHz frequency provides
a stronger LoRa signal due to greater SNR and RSSI values.
Nevertheless, 868 MHz ISM band shows more stable packet
(b) reception at every spreading factor. Considering that the less
the SF value is the less energy is being consumed [17], it
Fig. 8: Percentage of received packets per SF at 433 MHz (a)
can be recommended to use 868 MHz LoRa modules at 10th
and 868 MHz (b)
spreading factor in nine-story buildings. With greater floor
Packet received percentage line graphs reflect that signal difference between modules it would be better to use higher
becomes more unstable at low spreading factors. values of spreading factor. In future studies 915 MHz
Transmission stability can be classified according to the frequency can be explored and all three LoRa ISM bands
following rules: more than 95 % of packets received means can be compared with each other indoors and outdoors.
a stable signal, less than 95 % but more than 80 % of
The results can be useful in development and
packets received means a partly stable signal, while less
configuring LoRa indoor networks. These results were used
than 80 % of packets received means an unstable signal.
to deploy a research infrastructure of the Internet of Things
based on LoRa at the National Research University Higher
School of Economics.
TABLE VII. CLASSIFICATION OF TRANSMISSION
AT DIFFERENT SF AND FREQUENCIES
REFERENCES
ISM SF Lowest Classification [1] Petäjäjärvi, J., Mikhaylov, K., Pettissalo, M., Janhunen, J., & Iinatti, J.
(2017). Performance of a low-power wide-area network based on LoRa
band percent technology: Doppler robustness, scalability, and coverage.
International Journal of Distributed Sensor Networks. DOI:
7 32.2% Unstable 10.1177/1550147717699412.
[2] Islam, B., Islam, M.T., Kaur, J., & Nirjon, S. (2019). LoRaIn: Making a
8 75.1% Unstable Case for LoRa in Indoor Localization. 2019 IEEE International
Conference on Pervasive Computing and Communications Workshops
(PerCom Workshops), 423-426. DOI: 10.3390/ijgi7110440.
433 MHz 9 88.2% Partly stable [3] Petajajarvi, Juha & Mikhaylov, Konstantin & Hämäläinen, Matti &
Iinatti, Jari. (2016). Evaluation of LoRa LPWAN technology for
10 93.2% Partly Stable remote health and wellbeing monitoring. 1-5. DOI:
10.1109/ISMICT.2016.7498898.
[4] D. Kutuzov, A. Osovskiy, O. Stukach, "Modeling of Interconnection
11 98.4% Stable Process in the Parallel Spatial Switching Systems". 2016 International
Siberian Conference on Control and Communications, SIBCON;
12 98.8% Stable National Research University "Higher School of Economics".
Moscow; Russian Federation; 12–14 May 2016; DOI:
10.1109/SIBCON.2016.7491852.
7 79.4% Unstable [5] Haxhibeqiri, J., Karaağaç, A., Van den Abeele, F., Joseph, W.,
Moerman, I., & Hoebeke, J. (2017). LoRa indoor coverage and
8 87.4% Partly stable performance in an industrial environment : case study. 2017 22ND
IEEE INTERNATIONAL CONFERENCE ON EMERGING
TECHNOLOGIES AND FACTORY AUTOMATION (ETFA) (pp. 1–
868 MHz 9 92.8% Partly Stable 8). Presented at the 22nd IEEE International Conference on Emerging
Technologies and Factory Automation (ETFA).
10 99.2% Stable DOI:10.1109/ETFA.2017.8247601.
[6] Hosseinzadeh, S., Larijani, H., Curtis, K., Wixted, A., & Amini, A.
11 99.4% Stable (2017). Empirical propagation performance evaluation of LoRa for
indoor environment. 2017 IEEE 15th International Conference on
Industrial Informatics (INDIN), 26-31. DOI:
12 99.4% Stable 10.1109/INDIN.2017.8104741.
[7] Dambal, Vageesh & Mohadikar, Sameer & Kumbhar, Abhaykumar &
Guvenc, Ismail. (2019). Improving LoRa Signal Coverage in Urban
and Sub-Urban Environments with UAVs. DOI:
Signal stability classification illustrates that, firstly, the 10.1109/IWAT.2019.8730598.
best packet reception occurs at higher spreading factors [8] Yousuf, A.M., Rochester, E.M., & Ghaderi, M. (2018). A low-cost
(from 10 to 12). Secondly, 868 MHz frequency provides, in LoRaWAN testbed for IoT: Implementation and measurements. 2018
IEEE 4th World Forum on Internet of Things (WF-IoT), 361-366.
DOI:10.1109/WF-IoT.2018.8355180.
2020 Moscow Workshop on Electronic and Networking Technologies (MWENT)
[9] D. Kutuzov, A. Osovsky, O. Stukach, D. Starov, "CPN-based model of [13] Ameloot, T., Torre, P.V., & Rogier, H. (2018). LoRa Indoor
parallel matrix switchboard". 2018 Moscow Workshop on Electronic Performance: an Office Environment Case Study. 2018 International
and Networking Technologies (MWENT), 14-16 March 2018, Applied Computational Electromagnetics Society Symposium - China
Moscow, Russia. DOI: 10.1109/MWENT.2018.8337180. (ACES), 1-2. DOI:10.23919/acess.2018.8669294.
[10] Erbati, M.M., Schiele, G.A., & Batke, G. (2018). Analysis of [14] Sagir, S., Kaya, I., Sisman, C., Baltaci, Y., & Unal, S. (2019).
LoRaWAN technology in an Outdoor and an Indoor Scenario in Evaluation of Low-Power Long Distance Radio Communication in
Duisburg-Germany. 2018 3rd International Conference on Computer Urban Areas: LoRa and Impact of Spreading Factor. 2019 Seventh
and Communication Systems (ICCCS), 273-277. International Conference on Digital Information Processing and
DOI:10.1109/ccoms.2018.8463224. Communications (ICDIPC), 68-71.
[11] Zhu, G., Liao, C., Suzuki, M., Narusue, Y., & Morikawa, H. (2018). DOI:10.1109/ICDIPC.2019.8723666.
Evaluation of LoRa receiver performance under co-technology [15] Heltec Automation. Available online: www.heltec.org (accessed on 27
interference. 2018 15th IEEE Annual Consumer Communications & September 2019).
Networking Conference (CCNC), 1-7. [16] Teach, Learn, and Make with Raspberry Pi – Raspberry Pi. Available
DOI:10.1109/CCNC.2018.8319183. online: www.raspberrypi.org (accessed on 27 September 2019).
[12] Eremin, V & Borisov, A. (2018). A research of the propagation of [17] Bouguera, T.; Diouris, J.-F.; Chaillout, J.-J.; Jaouadi, R.; Andrieux, G.
LoRa signals at 433 and 868 MHz in difficult urban conditions. IOP Energy Consumption Model for Sensor Nodes Based on LoRa and
Conference Series: Materials Science and Engineering. 363. 012014. LoRaWAN. Sensors 2018, 18, 2104. DOI: 10.3390/s18072104.
DOI:10.1088/1757-899X/363/1/012014.