2

A Survey on LoRa Networking: Research Problems,

Current Solutions and Open Issues
Jothi Prasanna Shanmuga Sundaram, Student member, IEEE, Wan Du, Member, IEEE,
Zhiwei Zhao, Member, IEEE

Abstract— Wireless networks have been widely deployed for cate small payloads to convey interesting data like humidity,
many Internet-of-Things (IoT) applications, like smart cities temperature and other variables over a longer distance in a
arXiv:1908.10195v1 [eess.SP] 24 Aug 2019

and precision agriculture. Low Power Wide Area Networking multi-hop manner.
(LPWAN) is an emerging IoT networking paradigm to meet
three key requirements of IoT applications, i.e., low cost, large The above requirements have led to a new branch of IoT net-
scale deployment and high energy efficiency. Among all available working technology, called Low Power Wide Area Networking
LPWAN technologies, LoRa networking has attracted much (LPWAN), as conventional IoT networking technologies like
attention from both academia and industry, since it specifies Zigbee and Bluetooth can only provide a shorter range [2],
an open standard and allows us to build autonomous LPWAN [7], [8]. LPWAN employs simple network topology and long
networks without any third-party infrastructure. Many LoRa
networks have been developed recently, e.g., managing solar distance communication with low data rates to attain high
plants in Carson City, Nevada, USA and power monitoring energy efficiency [9]. Existing LPWAN technologies can be
in Lyon and Grenoble, France. However, there are still many divided into three categories i.e., networks based on cellular
research challenges to develop practical LoRa networks, e.g., infrastructure [10], [11], networks using third-party infrastruc-
link coordination, resource allocation, reliable transmissions and ture [12], autonomous LPWAN networks without any third-
security. This article provides a comprehensive survey on LoRa
networks, including the technical challenges of deploying LoRa party infrastructure [13].
networks and recent solutions. Based on our detailed analysis First, existing cellular technology covers a wide area but its
of current solutions, some open issues of LoRa networking are energy efficiency does not match LPWAN requirements as they
discussed. The goal of this survey paper is to inspire more works were not commissioned for machine-type communications.
on improving the performance of LoRa networks and enabling As cellular networks are already densely populated, a new
more practical deployments.
massive wave of IoT devices cannot be handled as it leads to
Index Terms—The Internet-of-Things, Low Powered Wide heavy interference. To overcome these challenges, intensive
Area Networking, LoRa, taxonomy. research is being conducted on Cellular-IoT technologies like
LTE-M [14], [15], NB-IoT [10], [11] and EC-GSM [11]. For
I. I NTRODUCTION example, NB-IoT [10], [11] operates at licensed Long-Term
Evolution (LTE) bands using Single-Carrier Frequency Divi-
T HE Internet-of-Things (IoT) applications [1], [2], like
smart homes and smart cities, become more and more
pervasive, which result in increasing density and scale of
sion Multiple Access (SC-FDMA) for uplink and Orthogonal
Frequency Division Multiple Access (OFDMA) for downlink
networked sensor deployments [3]–[5]. Ericsson mobility re- modulation. It facilitates higher Quality-of-Service (QoS) [16].
port [6] states that connected IoT devices will grow from Second, some service providers, like SigFox [12], In-
seven billion in 2017 to 20 billion in 2023, corresponding to genu [17] and Weightless [18], are proprietary networks.
an annual growth rate of 19%. The IoT applications employ Ingenu [17] is a founding member of IEEE 802.15.4K task
things with sensing capabilities to sense the environment, group. It leans on completing its stack whereas SigFox and
communicate with other devices and humans and make intelli- LoRa focus on faster time to market. It operates at the 2.4GHz
gent decisions. To connect IoT devices, wireless networks are band. Ingenu uses Random Phase Multiple Access (RPMA)
required to provide robust operations and wider coverage with modulation which gives higher link budget and coverage
high energy efficiency [1]. The IoT end devices are mostly while energy efficiency becomes a downside. Ingenu also
battery powered. They are expected to work for a longer span suffers from interference of other technologies like WiFi, low
of five to ten years without any maintenance. These IoT end- structural penetration of signals and increased propagation
devices are also expected to cover a large geographical area. loss at high frequencies [19]. SigFox [12] is more popular
For example, the forest monitoring application installs end in European region because of the traction made by widely
devices throughout the forest region. The devices communi- available vendors like Axom, Texas Instruments and Silicon
labs. 100Hz bandwidth (BW) and Ultra-Narrow Band technol-
ogy are utilized for transmitting smaller packets (12 bytes, up
Wan Du and Jothi Prasanna Shanmuga Sundaram are with the Department
of Computer Science and Engineering, the University of California, Merced. to 140 messages per day) at low data rates (up to 100 bits per
E-mail: {wdu3, jshanmugasundaram}@ucmerced.edu, Zhiwei Zhao is with second) modulated with Binary Phase Shift Keying (BPSK).
the School of Computer Science and Engineering, University of Electronic Major limitations of SigFox includes (i) being proprietary
Science and Technology of China. Email: zzw@uestc.edu.cn. Wan Du is
the first corresponding author of this article and Zhiwei Zhao is the second closed source technology, (ii) Low security mechanisms and
corresponding author. (iii) restrictions on downlink transmissions [20].
 3

TABLE I: Comparison of LPWAN technologies

LoRa Ingenu Sigfox NB-IoT

Third-party infrastructure Open source Closed source Closed source Open Source
Operating Band ISM Sub-GHz ISM 2.4GHz ISM Sub-GHz Licensed LTE band
180KHz
Channels Multiple SF with 64+8 40 1-MHz channels, 360 channels 3 DL and 2 UL
UL and 8 DL 1200 signals per second
Modulation CSS, FSK RPMA-DSSS, CDMA DBPSK, GFSK OFDMA, SC-FDMA
Data rate 0.3-37.5 Kbps 78 Kbps UL, 19.5 Kbps 100 bps UL, 600 bps DL up to 250kbps
DL
Communication Range 5Km -15Km [9], [10] 15Km [9], [10] 1-Km to 5-Km [9], [10] up to 35Km [9], [10]
Payload Length up to 250Bytes 10 Kilobytes 12Bytes UL and 8B DL 1600 Bytes
Authentication Symmetrical Mutual Authentication Burnt-in symmetrical Mutual Authentication
Authentication key authentication key
Encryption AES 128bit AES 256bit 5 LTE encryption

Finally, LoRa networking [21] is widely used for LPWAN TABLE II: Acronyms found in this paper
applications because, LoRa networking is an open-source Acronym Description
technology that enables autonomous network set-up at low bps bits per second
cost. LoRa networks have been widely deployed for many CDMA Code Division Multiple Access
applications and research systems. The openness of LoRa CR Code Rate
makes it an excellent choice for diverse IoT deployments [13]. CSS chirps Spread Spectrum
General IoT applications include smart buildings [22], smart DBPSK Differential Binary Phase Shift Keying
cities [23], smart agriculture [24], smart meters [25], [26] and DL Downlink
water quality measurement [27]–[30]. DSSS Direct Sequence Spread Spectrum
EC-GSM Extended Coverage-GSM
Major LPWAN technologies are compared in Table I.
FSK Frequency Shift Keying
Acronyms used in this table are described in Table II. Working GFSK Gaussian Frequency Shift Keying
in sub-GHz band using CSS modulation makes LoRa tech- Kbps Kilo bits per second
nology immune to interference as the chirp signal varies its LPWAN Low Power Wide Area Networks
frequency linearly with time. The chirp signals utilize the NB-IoT Narrow Band-IoT
available bandwidth instantaneously consuming low power OFDM Orthogonal Frequency Division Multiple Access
than the other LPWAN technologies. A nominal coverage RPMA Random Phase Multiple Access
of 5Km-15Km [10] is obtained with higher payload (up SC-FDMA Single Carrier-Frequency Division Multiple Access
to 250Bytes) when compared to other technologies. LoRa SF Spreading Factor
networks offer better downlink capabilities than Sigfox and UL Uplink
Ingenu. LoRa networking provides light-weight encryption
and authentication mechanisms that can be configured during
activation. Another important advantage of LoRa networks is col, QoS, latency, communication range and deployment cost
that the configuration and firmware updates can be sent over of each technology [16]. LPWAN application scenarios are
the air [31]. categorized and some important parameters to be considered
Why a new survey on LoRa networking? Raza et. al for each specific scenario are studied. Research challenges
ascertain the need of LPWAN by justifying the inability of and recent technical advancements of each technology are not
legacy wireless systems to comply with the constraints of discussed in detail.
LPWAN [9]. The design goals of LPWAN along with various Different from the above mentioned surveys [9], [16], our
techniques to achieve these goals are discussed. On discussing survey is focused on LoRa networks. We study the recent
the challenges and research directions, the authors find that performance measurements of LoRa networking [26], [32]–
most of the working groups focus on PHY and MAC layers. [36], [36]–[50] to understand and devise a taxonomy for
We argue that the upper layers should also be discussed such as the research problems of LoRa networking. The recent so-
the efficient deployment of LPWAN. A brief description is pro- lutions [21], [41], [51]–[76] are further discussed in detail
vided on technical specifications of all LPWAN technologies to understand the advancements of LoRa technology. Finally,
while recent performance measurements, research challenges we present some open issues that could further improve the
and solutions are not discussed in detail. performance of LoRa networking.
Sinha et al. study two leading LPWAN technologies, LoRa A survey on LoRa technology has been recently pub-
and NB-IoT, by comparing their physical features, MAC proto- lished [77]. It discussed the literature, solutions and open
 4

Fig. 2: Class A receive windows

Fig. 1: LPWAN Network architecture windows (RW)”. Figure 2 depicts the RWs of Class A op-
erating mode. Frequency and data rate of the first RW is
the same as the uplink transmission parameters whereas the
issues without any classifications. In our article, a clear taxon- second slot operates on pre-agreed parameters to improve
omy has been devised. Based on this taxonomy, the challenges, the robustness of message transmissions. end-devices do not
current solutions and open issues are discussed with tabulated expect replies from the server for unconfirmed messages. Class
version of system analysis and hardware experiments. The tax- B operating mode opens additional receive windows scheduled
onomy provided in this article facilitates a clear understanding by gateways through beacon packets. Class C mode has no
of the challenges, solutions and open issues. In addition, some downlink restrictions and can receive downlink messages any
of the recent contributions to LoRa networking [51], [52], [78], time whenever it is not in a transmitting state.
[79] are also discussed in our article. In general, LoRa denotes the physical layer while Lo-
The rest of this paper is structured as follows. Section II RaWAN denotes the MAC layer communications and network-
gives a brief description of LoRa technology. Section III ing in LoRa stack.
lists the existing deployments of LoRa networks and their LoRa. The physical layer of LoRa technology uses Chirp
advantages. Section IV investigates the research challenges Spread Spectrum (CSS). Chirps are the signals whose fre-
of LoRa networking and devises a taxonomy. Section V quency varies linearly with time within the available band-
gives a comprehensive study on how these research problems width. This attribute makes the chirp signals resilient to noise,
are tackled by recent solutions. Section VI discusses some fading and interference. Every LoRaWAN packet starts with a
open issues that still needs to be addressed and Section VII preamble of ten chirps and six synchronization chirps followed
concludes the article. by the data. Each chirp can modulate multiple chips (data
bits). The number of data bits modulated depends on the
II. A B RIEF TECHNICAL BACKGROUND OF L O R A parameter Spreading Factor (SF). For example, nine bits can
be encoded in a chirp using SF9. A message sent with higher
This section briefly describes the technical features of LoRa. SF takes more time on air and reduces the data rate but
LoRa operates in unlicensed sub-GHz ISM band (900MHz improves resilience to noise. LoRa modulation also has two
in USA and 860MHz in Europe). Using 125KHz, 250KHz other parameters namely Bandwidth and Code Rate (CR). The
and 500KHz of bandwidth, smaller payloads of up to 250 bandwidth can be set to 125KHz, 250KHz and 500KHz and
bytes can be transmitted over a distance of 5-15 Km and the CR can be set to 4/5, 4/6, 4/7 or 4/8.
the system can last up to 5-10 years consuming low power LoRaWAN. The LoRa community often refers LoRaWAN
according to the recent report [10]. A LoRa system comprises as a “MAC in the cloud” design [52]. Gateways are the
of end-devices, gateways, network and application servers. forwarders acting based on commands from the servers. All
Figure 1 depicts the architecture of a typical LoRa system. MAC decisions like data-rate, handling ACKs are decided at
end-devices collect information and send them to Gateways. the servers. LoRaWAN MAC employs two modes to divide
Gateways relay messages between end-devices and network air-time between end-devices for handling collisions. The first
servers. A network server is configured to direct messages mode is the ALOHA MAC that allows end-devices to transmit
to appropriate application servers for processing. There as soon as they wakeup and exponential back-off is applied in
are three operating modes for LoRa. LoRa end-devices must case of collisions. The second mode is the TDMA scheduler
implement Class A operating mode. Other optional modes where the network server allots time-slot for each end-device
like Class B and Class C can also be utilized. The end- to transmit their messages.
devices operating in Class A and Class B modes are generally
battery powered while the end-devices operating in Class C
is mains powered. Class A utilizes less energy than Class B A. Unique properties of LoRa
and C. In Class A, after sending confirmed messages, end- LoRa technology has some unique properties making it a
devices expect an acknowledgment (ACK) from the Network widely used technology. The unique properties are (i) Ultra-
server during two pre-agreed time-slots known as “receive long distance, (ii) Low cost and complexity devices (iii)
 5

Long lifetime of nodes, (iv) concurrent reception capacity of TABLE III: In-field deployments of LoRa networks
gateways and (v) robustness in Doppler effect. All these unique LoRa Deployment Location
properties are experimentally verified by [80]. Waste Management [23] Seoul, North Korea
Ultra-Long distance: In Line-Of-Sight (LoS) communi- Solar power plant management [23] Nevada, USA
cations, the longest SF12 can achieve a distance of up to Power usage monitoring [23] Lyon, France
9Km with Packet Reception Ratio (PRR) > 70% and the Power usage monitoring [23] Grenoble, France
smallest SF7 can achieve a distance of 5Km for PRR > Smart meters [25] Gehrden, Germany
70%, according to the report in [34], [60]. In Non-Line-of- Smart golf course [83] Calgary, Canada
Sight (NLoS) scenarios comprising of buildings, the longest Smart Islands [84] Mallaorca, Spain
distance achieved is around 2Km [79]. It is also noticable
that the communication distance is affected by the parameters
Bandwidth, SF, transmission power and coding rate [40]. the capacity of waste-bins became tedious. City management
Low cost and complexity: The LoRa devices are fabricated installed LoRa-enabled smart bins to periodically collect the
such that they are not complicated hence reducing the price. capacity of waste-bins. This helped to clear the bins as soon
Reduced complexity also reduces the overheads incurred dur- as they are filled. This application gives a 66% reduction in
ing communications. For example, a sophisticated CSMA is waste collection frequency, an 83% reduction in costs and a
not employed instead a CAD is employed that will just check 46% increase in recycling.
for preambles in the channel before transmission. There is Carson City management in Nevada found that effective
no signalling overhead like other traditional communication transition between legacy and solar power is important when
networks. Whenever a node wants to transmit, it wakes up, solar efficiency reduces during cloudy climates and nights.
checks for channel status, transmits and goes back to sleep. LoRa-enabled monitoring system monitors the current envi-
Long lifetime: The LoRa consumes around 120-150 mW ronment status of the solar-deployed site. Decision to use
during transmission and 10-15 mW for MCU operations solar or legacy power is based on this collected data. This
based on different radios and host-boards used. This can be system reduced 15% of operational expenses and boosted solar
extrapolated to 2-5 years in total life time while the duty cycle power output to 75,000 kWh of clean power because of proper
is varied from 0.1% to 10% [80]. transition between solar and legacy power.
Concurrent reception of gateways: Current LoRa gate- In Lyon and Grenoble, power consumption monitors were
ways are capable of concurrent reception on 8 channels. Even deployed in households. This helped the residents to monitor
the same SF can be received on different channels. All the their power usage and turn-off unwanted devices. This system
different spreading factors from SF7 - SF12 are orthogonal helped to reduce power consumption by 16%.
and transmissions with different SFs can be received on the Smart meters: Semtech’s white paper [25] describes and
same channel concurrently. evaluates the capacity of LoRa technology for smart metering
Robustness to Doppler effect: Liando et al. [80] prove applications. This system is deployed in Gehrden, Germany
that LoRa signals are robust to Doppler effect. The CSS where the population is 15,000. Around 7000 households were
modulation used by LoRa is highly resistant to Doppler effect. installed with LoRa-enabled smart meters. This application
Mobile LoRa end-devices at a constant speed and in LoS can helped to reduce the man-power utilized for monitoring power
yield PRR > 85% [80]. usage by transmitting meter-readings periodically to the gate-
way.
III. E XISTING DEPLOYMENTS OF L O R A NETWORKS Smart golf course: Shaganappi point is a popular golf
This section explains the existing deployments of LoRa course in the city of Calgary, Canada serving a community of
networking and their advantages. There are many use cases million people. Golfers play an average of 90,000 - 100,000
like building management system [22], smart agriculture [24], rounds of golf during April - November every year. It is identi-
smart parking [81] and smart lighting [82]. some popularly fied that slow play devalues the overall experience. Improving
known real-world deployments are summarized in Table III. this experience will help to retain customers. Hence, each
An overview of these deployments is discussed below. golf cart is fixed with a LoRaWAN sensor. With real time
Smart cities and Urban Deployments: Influx of popu- movement and location of golf carts, pause of play anomalies
lation toward cities demands a better way of governing and are detected and appropriate help is provided to speed up the
organizing amenities for optimal usage. Semtech’s white paper play. Large coverage of golf course requiring periodic updates
[23] explains how LoRa LPWAN could provide efficient usage with low power consumption makes LoRa a perfect solution
and governance to make cities smart. Some applications that for this use case. Hence overall experience of customer is
could improve daily life of the people are Smart parking [81], elevated with maximization of revenue [83].
Smart lighting [82]. These applications will improve people’s Smart Islands: Mallorca is largest of Spain’s five Balearic
living experience. Some in-field deployments [23] are (i) waste Islands popular for white sand and turquoise water. Currently
management in Seoul, North Korea, (ii) integrated sensing of there are 25 people for every meter of beach on the Island with
Solar power plants in Carson city, Nevada, USA, (iii) power 32% anticipated growth by 2030. Citizens have shifted their
monitoring in Lyon and Grenoble, France. attitude to conserve natural resources in the Island. LoRaWAN
Seoul experiences humongous floating crowds every day. sensors are installed to aid water management systems. This
As the crowd moves through the city dynamically, probing periodically reports water quality and levels [84]. This system
 6

has seen 25% water savings since installation. spectrum. Possibilities for these end-devices transmitting data
concurrently varies based on the application. Multiple access
IV. TAXONOMY OF RESEARCH PROBLEMS is to allow multiple end-devices to share the limited spectrum
In this section, we study the research challenges of LoRa for communication. The multiple access problems involve
networking. The severity of these challenges is identified by two different aspects namely Link coordination and Resource
investigating its effect on the operations of LoRa technology. allocation.
Finally, a taxonomy is devised to categorize these challenges. Link coordination: Deploying thousands of devices require
Figure 3 illustrates the taxonomy of the challenges of LoRa multiple-access to improve concurrent transmissions and avoid
networking. collisions. Links are coordinated through MAC protocols. The
LoRa networking employs ALOHA and TDMA scheduler to
coordinate links. These techniques cannot handle collisions
A. Energy Consumption
while thousands of devices are connected to the network
The most important characteristic of LPWAN is its high [42], [43]. Thus, new techniques for handling collisions and
energy efficiency. This becomes an important parameter to coordinating links are required. This will help to upscale the
improve the longevity of end-devices. LoRa networks are density of LoRa deployments [44], [45].
expected to work for a longer period of 5-10 years with Resource Allocation: In LoRa technology, the transmission
minimal maintenance. Hence, power consumption becomes a is controlled by the parameters Spreading Factor (SF), Trans-
major challenge for LoRa networking. end-device operations mission power (TP), Bandwidth (BW) and Channel. Varying
can be classified into (i) micro-controller operations and (ii) these parameters result in different transmission qualities.
wireless transmissions. Power consumed for micro-controller This can be leveraged to improve concurrent transmissions.
operations vary according to the chosen host board but the Dynamically allocating reasonable resources to end-devices
power consumed for wireless transmissions completely de- based on the deployed environment improves multiple access
pends on the LoRa technology. Charm [52] shows that wireless and thus scalability [85].
transmission extorts more power than micro-controller oper-
ations. LoRa technology employs two techniques to reduce
energy consumption, (i) consuming instantaneous bandwidth
for transmitting a chirp signal and (ii) not employing heavy D. Error Correction
MAC protocols for scheduling. In spite of these techniques,
end-devices consume more power than expected due to some LoRa technology communicates data over long distances.
unavoidable circumstances like retransmissions caused by While the message is transmitted over the air, it is possible
channel impairments. for the data to get corrupted or lost due to channel effects,
environmental conditions or collisions. Existing error correc-
tion schemes of LoRa networking, like hamming code, cannot
B. Communication Range
aid data corruption or loss efficiently [56]. LoRa technology
Large communication range is also an important rudiment offers different spreading factors to make the signal more
of LoRa technology. Current LoRa technology relies on chirps resilient to noises. SF12 is the stronger spreading factor but
spread spectrum, which is more resilient to interference. LoRa it takes more time-on-air. Even these signals can also be
networking will be deployed in many scenarios such as homes, corrupted in dense environments [33], [34]. There are two
hospitals, schools, forest, etc. end-devices will be placed in the types of current solutions namely (i) channel coding and (ii)
locations open to air, closed by concrete or steel, etc. Aiming interference cancellation.
such diverse deployment conditions, signal attenuation, propa- Channel coding. A recent channel coding technique pro-
gation losses and fading have to be countered to improve signal posed is DaRe [56]. DaRe is an application layer coding to
penetration thus improving the coverage of LoRa networks retrieve lost data using redundant data. Sandell et al. [59]
[35], [32]. It has been noted that gateways can detect signals explain that this technique cannot aid bursty packet loss and
below a given threshold but cannot decode them. Devising a has some limitations that bounds the performance.
technique to decode these signals will improve communication
Interference Cancellation: Even though channel coding
range. Another important challenge is estimating the coverage
aids error, they cannot guarantee error correction in the case
of LoRa networks. Chall et al. [79] study different models
of collided signals. Interference Cancellation will extend error
through empirical measurements in Lebanon. Demetri et al.
correction by untangling and extracting data from collided
[78] identified that LoRa’s signal coverage is anisotropic. This
signals. Recently proposed technique Choir [51] and Netscatter
is because LoRa signals travel a longer distance and experience
[55] cancels interference but Choir’s [51] limitation is pointed
varying environments with dynamic and static obstacles in
out by Netscatter [55]. Netscatter [55] can cancel collisions
different directions. Mathematical models and systems for link
of 256 concurrent transmissions, which is not adequate to
quality estimation are still unexplored for LoRa.
handle thousands of end-devices by a single gateway. For
example, a large-scale temperature monitoring system requires
C. Multiple Access: all end-devices to transmit data at the same time [25]. Hence,
LoRa networking aims to connect thousands of end-devices Interference cancellation is still an open problem that should
to the network, communicating over a confined region and be addressed to improve the performance of LoRa Networking.
 7

Fig. 3: Taxonomy of LoRa Research Challenges

E. Security to 2.8 Km while consuming only 9.25 µwatts of power at the

For all computer communications, security is a major con- rate of 37.5 Kbps. Power consumption is reduced by nearly
cern. There are many security attacks like eavesdropping, 1000× than standard LoRa technology. These passive RF chips
selective forwarding and node impersonation [86]. All the can be powered through solar panels attached to them. This
above mentioned attacks try to obtain the key used for technique is also analyzed over home/office sensing, precision-
encryption. If this key is compromised, the entire system sensing of agriculture devices and epidermal devices to prove
can be broken. Currently, LoRa technology uses a symmetric their efficiency.
key cryptographic technique with AES-128 bit encryption. PLoRa [53] proposes a hardware and software co-design
Existing LoRa technology generates the key and never updates to enable battery-free LoRa networks, operating on the en-
it. Hence, Key generation and Key update mechanism is a ergy harvested from solar devices. The proposed PLoRa tag
major concern. Third-party authorization is required when the transmits data by backscattering ambient LoRa transmissions
application and network service providers are different. So, without external excitation signals unlike LoRea [88] and
these applications require third-party authorization to ensure LoRa Backscatter [54] which uses dedicated hardware for
privacy and security [47], [75], [76]. generating excitation signals. The active LoRa signal emitted
by a gateway or a node is converted into Passive LoRa signals
V. C URRENT S OLUTIONS to send data using ON-OFF keying technique. The power
consumption of PLoRa is 250× smaller than the standard LoRa
In this section, we summarize the recent measurements technology.
and current solutions proposed to address the problem power
Charm [52] improves the battery lifetime up to 4x the stan-
consumption, communication range, error correction, multiple
dard LoRa technology by avoiding retransmission of the weak
access and security.
signals. This technique is discussed in subsection V-B. FADR
Existing works can be classified into two categories as [62] reduces the power consumption of standard LoRaWAN by
performance measurements and current solutions tackling the 22%. Wireless power charging has been a promising solution
above challenges. All the recent experiments, measurements to handle energy consumption problem for wireless sensor
and simulations use Class A end-devices unless specified. networks [89]. Realizing them on low-cost LoRa hardware
Table IV summarizes performance measurements and case has been done in [90]. A circuit has been designed to enable
studies conducted on LoRa. Table IV differentiates the mea- wireless power transfer on LoRa enabled sensor nodes.
surements made through theoretical analysis, test-bed and
simulated evaluations. It is to be noted that no performance Gao et al. [91] investigated energy fairness problems in
measurements have been made on power consumption and LoRa networks. Due to large differences between data rates
error correction capabilities. Table V enumerates recent solu- used by different end-devices, end-devices far away from the
tions to improve the performance of LoRa networking. The gateway have to use a low data rate and spend more energy to
current solutions addressing more than one problem can also transmit certain amount of data. To make energy consumption
be identified in Table V. more fair across all end-devices in a LoRa network, Gao et al.
[91] propose to deploy more gateways to allow end-devices
to use high data rates to reach at least one gateway. To make
A. Energy Consumption the network transmission more efficient, a network model is
Various techniques are employed to improve battery lifetime developed and used to allocate network resource to each end-
of the end-devices. Some works propose techniques (i) to device. A heuristic search based network resource allocation
harvest ambient energy from the environment [87], [54], [53]; algorithm is developed to find the best network setting for each
(ii) to use backscatter signals for transmission, [54], [53], and end-devices.
(iii) to detect and decode weak signals and increase data rate Key Insights. Ambient energy harvesting is used to make
to reduce power consumption [52]. LoRa devices battery free [53]. One of the most power con-
LoRa Backscatter [54] proposes a backscatter system for suming operation in LoRa wireless transmissions is generating
LoRa based on CSS modulation. Data can be transmitted up carrier signals [52]. One way to reduce this part of power
 8

TABLE IV: Performance measurements of LoRa

Articles Communication Range Packet Delivery Multiplexing Security

T/M S T T/M S T T/M S T T/M S T
Harris et al. [35]
Fuidiak et al. [32]
Petajajarvi et al. [33]
Petajajarvi et al. [34]
Haxhibeqiri et al. [36]
Mikhaylov et al. [37]
Magrin et al. [26]
Semtech White paper [38]
Bankov et al. [39]
Haxhibeqiri et al. [36]
Angrisani et al. [40]
Blenn et al. [41]
Lauridsen et al. [42]
Vejlgaard et al. [43]
Zhu et al. [44]
Orfanidis et al. [45]
Ferre et al. [46]
Aras et al. [47]
Butun et al. [48]
Miller et al. [49]
Oniga et al. [50]
Liando et al. [80]
Table notes
T/M - Theoretical / Mathematical Analysis
S - Simulated evaluation
T - Testbed evaluation

consumption is to utilize the backscatter signals [54]. Another 1,000,000 packets at the rate of 417 packets per hour. For
method is to leverage passive chips [92] for carrier generation. all the three phases, the achieved delivery rate is more than
Charm [52] identifies that LoRa is able to receive weak signals 95%. This paper notes that multiple gateways will scale-up the
but not able to decode them. Wireless power transfer [89] is network and improve communication range as end-devices can
not feasible owing to the complex, costly hardware extensions communicate with more than one gateway.
in low power, low cost LoRa modules. Placing more gateways Navarro et al. [89] and Haxhibeqiri et al. [36] evaluated
and dynamic allocation of TP addresses energy consumption communication range of LoRa in Industrial environments.
problem and improves network lifetime. The industrial environment at Royal Flora Auction Center,
Netherlands, covering 250000m2 of both indoor and outdoor
B. Communication Range spaces. LORANK [90] Gateway was fixed 6 m above the floor
Testbed Measurements. Semtech’s white paper [38] evalu- and WiMOD-IM880A [93] end-devices were attached to the
ates the capacity of LoRaWAN in dense urban environments. trolleys 1.7 m above the ground. The nodes were triggered
Ten gateways operating on 8 channels were used for trials with to transmit at SF7 and SF12. Measurements were taken from
100 sensors as end-devices to transmit at different data rates to 43 measuring points covering indoor, outdoor spaces. Fifty
mimic like 10,000 end-devices. Three phases of experiments packets were sent from each test point. Measurements show
with varying quantity of packet transmission were conducted that a maximum of 6000 end-devices can be handled by a
to evaluate packet delivery. The first phase contains 250,000 single gateway with a packet loss rate less than 10%. Packet
volume of packets being transmitted at the rate of 104 packets loss is around 6% when less than 3500 end-devices are used.
per hour. The second phase generates 500,000 packets at the Petajajarvi et al. [33] analyse the range of LoRaWAN with
rate of 209 packets per hour and the third phase generates 14 dBm Transmit Power (TP) and the largest SF at Oulu,
 9

TABLE V: Summary of recent solutions for LoRa Challenges

Article Multiplexing Power Consumption Communication Range Error Correction Security

Choir [51]
Charm [52]
LoRa Backscatter [54]
PLoRa [53]
Netscatter [55]
DaRe [56]
Blenn et al. [41]
Bor et al. [21]
Chall et al. [79]
Demetri et al. [78]
Georgiou et al. [57]
Donmez et al. [58]
Sandell et al. [59]
Haxhibeqiri et al. [60]
Reynders et al. [61]
Abdeel et al. [62]
Reynders et al. [63]
Pop et al. [64]
Cuomo et al. [65]
Van et al. [66]
Mikhaylov et al. [67]
Cattani et al. [68]
Voigt et al. [69]
Lee et al. [70]
Kim et al. [71]
Na et al. [72]
Girard [73]
Kim et al. [74]
Naoui et al. [75]
Liando et al. [80]

Finland. The experiment is conducted for 14 days during coverage. Maximum range noted on ground and water is 15
spring and summer. The population of Oulu is around 200,000 Km and 30 Km respectively. Total packet loss ratio for on-
people with high rise buildings. Throughout the experiment, ground measurements and on-water measurements become
Kerlinks LoRa gateway [94] is fixed on a tower 24 m above 34% and 32% respectively. With these measured data, channel
the sea level with -137 dBm sensitivity in order to find the attenuation model is calculated for areas similar to Oulu.
maximum communication range. Semtech 1272 transceiver
[95] is used as an end-device. For on-ground measurements, Research solutions. Du et al. [96], [97] proposed a solution
end-devices are fixed on a car’s roof-rack, approximately 2 to improve the communication range of sparse wireless sensor
m above the ground, which drove around major cities at 40 networks. Choir [51] identifies an intrinsic property of LoRa
Km/h - 100 Km/h. For on-water measurements, end-devices radios in which the carrier frequency varies by a small bound
are fixed on a radio mast of the boat. end-devices send packets (902.4 MHz instead of 902.7 MHz) because of cheap radios.
periodically including GPS coordinates. The SF of end-devices This is exploited to disentangle collided signals and extend
is set to SF12 because the goal was to find the highest possible the range up to 2.64× the standard LoRa technology. The
nodes that are far away from the gateway transmit signals
 10

whose SNR goes below the noise floor. It is assumed that the 1) Channel Coding: A new application layer data recovery
neighbouring nodes send data that do not vary to a greater technique called DaRe [56] is proposed based on Convolu-
extent. These physically co-located nodes, far away from the tional and Fountain codes. This technique extends data with
gateway, are coarsely synchronized through Class B beacons redundant information. These redundant data are chosen from
to transmit data at the same time to enable constructive the previous data units so that the lost frame can be calculated
interference thus improving the SNR above the noise floor. from the other received frames. The disadvantage is that
Gateways can receive and decode this collided signal to obtain previous data units should be buffered in the memory for
approximate data of the region far away from gateways. This computing redundant information. This makes the generator
technique improves the communication range by 2.64× the matrix banded thus inducing difficulty to create degree of dis-
standard LoRa technology. tribution according to LT codes. Sandell et. al [59] show that
Charm [52] proposes a new hardware and software co- the memory affects performance and complexity. While DaRe
design to extend coverage and battery life of LoRa devices. uses a complex Gaussian elimination making the decoding
This is achieved by allowing multiple gateways to send weak process complex, an optimized decoding technique is proposed
signals (that cannot be decoded by a single gateway) to cloud in [59].
and coherently combine them to decode data. Programmable Sandell et al. [59] analyses the technique proposed in [56]
auxiliary hardware attached to the gateway improves the and shows that for larger packet loss probabilities, reducing
gateway’s ability to detect very weak signals which cannot be the code rate increases the interference leading to reduced
directly detected by gateways. Joint decoding algorithm uses efficiency of data recovery. Showing the relationship between
a heuristic approach to select signals to be combined at the latency of decoding algorithm and packet loss probability, a
cloud. Results show that the range is improved 3× and battery less complex decoding algorithm like Accumulative Gaussian
life is improved 4× the standard LoRa technology. Elimination is proposed by Du et. al [97] to reduce latency.
LoRa Backscatter [54] discussed in subsection V-A can Finally, the paper concludes that data recovery through redun-
send data to a receiver located 2.8Km away. Sensor Networks dancy techniques increase number of transmissions and thus
Over Whitespaces (SNOW) [98] was first designed for sensor collision. So it cannot be used solely to aid packet loss.
networks to be connected over a wide area. As the traditional 2) Interference Cancellation: Choir [51] and FTrack [100]
sensor networks cannot communicate to a longer distance, propose novel solutions to disentangle and decode collided
TV whitepsaces that could communicate a long distance is signals. Using constructive interference, a sparse overview of
exploited. Scalability and energy efficiency is achieved by the data is obtained from a group of geographically co-located
splitting carrier into several sub-carriers with parallel packet end-devices far away from the gateway. Test bed evaluation
receptions. The PHY layer handles OFDM modulation and shows that throughput is improved 6.84× than the standard
the MAC layer handles sub-carrier allocation. This technique LoRa. It is also proved that Choir yields better results than
is extended by Rahman and Saifullah [99] to integrate multiple multiple antenna deployments.
LPWANs and improve the communication range specifically NetScatter [55] improves interference cancellation. It the-
in infrastructure-restricted rural areas. The nodes located far oretically proves that Choir [51] can only decode 5-10
away communicate with gateways over white spaces meant concurrently transmitting devices. A new distributed coding
for TV signal communication. This technique is implemented technique based on CSS is proposed to decode concurrent
on generic GNU radios. Its function on LoRa devices is still transmissions below noise floor in a single FFT operation.
ambiguous. Experimental results show that this technique can decode
Key Insights. The above measurement experiments throw 256 concurrent transmissions with 14-62× improvement in
light on (i) placing more gateways to improve network density, throughput and 15-67× improvement in latency when com-
coverage and reduce energy consumption (ii) less packet loss pared with existing interference cancellation techniques.
in networks with sparse end-device placements [33]. Besides, Key Insights. The key insights regarding error correction
we can infer the following insights: are listed as follows.
• Radio imperfections occur in LoRa due to cheap radios • An intrinsic property of LoRa to deviate from carrier fre-
put into use. quency is identified and exploited in Choir [51] to correct
• These radios generate slightly different carrier frequencies errors because of signal collision. Choir is analyzed and
than specified [51]. its inability to scale-up is identified and a new distributed
• LoRa gateways can receive weak signals but cannot error correction technique is proposed in Netscatter [55]
decode them [52]. for scalability.
• Extending coverage of LPWAN’s is experimented with • Recently proposed channel coding technique, DaRe [56],
GNU radios but its operation on commercial LoRa de- and its analysis [59] shows that there is a heavy potential
vices has not yet been studied [99]. for characterizing efficient channel coding techniques
for LoRa to improve error correction, thus improving
C. Error Correction reception rate and network lifetime.
In this subsection, the research solutions for error correction
are classified into channel coding and interference cancella- D. Multiple Access
tion. The solutions are explained in detail before concluding In this subsection, the solutions for multiple access are
with key insights. divided into Link coordination and Resource allocation. The
 11

measurements and solutions for each category are explained the outcomes of real-word test-bed and the results from other
below with their key insights summarized at the end of each works to prove the model’s accuracy. It can be seen that even
category. though CSMA cannot provide better results for dense nodes,
1) Link Coordination: Measurements. Bankov et. al [39] CSMA-10 is achieving better performance than CSMA. The
identifies four important issues. The first issue is whether the performance comparison of p-CSMA and CSMA is conducted
gateway should listen to the channel during the interval, T1, by Kouvelas et al. [102] for small scale networks expound that
between frame reception and transmission of response. There p-CSMA is an important step to be taken for improving the
arises a problem if the channel is busy with other scheduled scalability of LoRa networks but it has not been realized yet
transmissions in that specific interval T1. This might cause on real-world devices.
delay to the ACKs, leading to unwanted retransmissions. The There are many retransmission policies devised for wireless
proposed solution is to cancel the pending transmission that networks [103]–[106] and wireless sensor networks [107]–
may cause collisions at the end-device and transmit ACKs [109]. A joint retransmission scheme with compression and
on the downlink channel. The second issue occurs when channel coding is developed for single-hop networks with
two transmitted frames are overlapping in the same time energy constraints like LoRa [110]. Its energy efficiency
interval over different channels. Gateway would not be able has been theoretically evaluated. This retransmission scheme
to acknowledge both messages at the specified window with retransmits the last q failed data blocks along with a new data
single downlink. The third issue is the limited interval for block using compression and coding schemes. Although the
retransmission. Due to the above said factors, ACKs may take performance has been proved theoretically, its implementation
more time to reach an end-device which will increase the with the required computational resource on real-time testbeds
retransmission probability even for a successfully delivered will be more helpful to understand the practical efficiency for
message. The authors recommend increasing the delay interval real-world deployments.
or use exponential back off to counter this issue. The fourth Key Insights. Tailoring different transmission parameter
issue arises when there is no optimal policy to select the data combinations for different end-devices while considering net-
rate for downlink. Simulations conducted show that packet work density is proven to be capable of improving link coor-
error rate and packet loss ratio increase with traffic due to dination [61]. Liando et al. [80] leverage the fact that LoRa
improper link coordination. gateways can receive packets with different SFs and the same
Research Solutions. Reynders et al. [61] address Lo- frequency simultaneously to reveal that PRR can be doubled
RaWAN’s scalability and reliability through a novel MAC if the SF is randomized on detecting a preamble during
protocol, RS-LoRa. This technique works in two phases. The CSMA. It is theoretically proven that a retransmission scheme
gateway sends coarse-grained information of allowed TP and employing data compression and channel coding improves link
SF for each channel as Class B beacons in the first phase. coordination.
In the second phase, each end-device selects one parameter 2) Resource Allocation: Measurements. Semtech’s White
combination from the beacon that better suits the node. As- paper [25] evaluates the capacity of LoRa technology for smart
signing different SFs with different parameter combinations metering applications. This system is deployed in Gehrden,
helps to alleviate the Capture Effect. Reliability of network Germany. Around 7000 households were installed with LoRa-
performance is improved by decreasing the Packet Error Ratio enabled smart meters. 11 Kerlink-V1 gateways [94] were
up to 20% than the standard LoRaWAN. This technique proves mounted on rooftops with 30cm/70cm half wave dipole an-
its superiority to the standard LoRaWAN through NS-3. tenna. A simple meter protocol was employed for reading a
Based on measurements obtained through real-world ex- seven-digit register. The payload allocates one byte for status
periments, Haxhibeqiri et al. [60] build a simulation model and three bytes for each register. The downlink payload length
to study the scalability of a single cell LoRaWAN based on is fixed as 10 bytes. Meters are configured to send unconfirmed
interference. It is showed that LoRa physical layer is robust payload every 15 minutes and confirmed payload once a day.
and can send six times traffic than the pure Aloha with 125 Class-C end-devices are utilized for this experiment. 24-hour
KHz bandwidth. Based on end-device density and their data raw data is used for the experiments. It is shown that a gateway
rates given in [101], simulations are carried out to determine with average throughput can handle 470,000 messages per
node density for different IoT applications. day. It is demonstrated that the network can be scaled up
A series of experiments are conducted to identify the po- locally by adding gateways. It is also shown that the ADR
tential of Channel Activity Detection (CAD) and Ideal-CSMA algorithm improves the network capacity by adjusting the data
on a dense network of 50 nodes by Liando et al. [80]. The rate, frame repetition rate and channel allocation.
results show that Ideal-CSMA fails to provide high reception Petajajarvi et al. [34] conducted experiments in the Uni-
rate when nodes synchronously perform channel detection. versity of Oulu, Finland. LoRaMote [111] is used as end-
Hence, the authors devise a CSMA-CAD with four additional device for measurements. These end-devices were configured
preambles in the packet to achieve doubled PRR. Whenever a to send messages to the base station every 5 seconds with
preamble is detected, the transmitting SF is randomized and no ACKs, no retransmissions and no ADR. End-devices are
the channel is sensed again for transmission. configured to transmit at six different channels. The gateway
The CSMA and CSMA-x are simulated on NS-3. CSMA- is the same as used in [33]. The packet delivery was above
x works similar to CSMA but it senses the channel for the 96.7% and 95% when the end-device was static and mobile
gap of x ms. The results of this model are compared with respectively. Similar results have been observed in [38] and
 12

[26]. Measurements show that most of the campus is covered Cattani et al. [68] conduct experiments to understand the
by SF7 itself. Interesting results were obtained while varying effect of tuning PHY parameters and environmental factors on
physical parameters. Farthest location is not reached by SF7 LoRaWAN communication reliability and energy efficiency.
and BW125. However, 60% of the packets were correctly Experimental results show that, for end-devices far away from
received from the same point with SF7 and BW250. For power the gateway, Packet Reception Ratio of fastest PHY setting is
consumption evaluations, RN2483 based LoRa module was only 10% lower than the slowest setting. Hence, the authors
added to a sensor and actuator kit with a Keysight’s power recommend selecting high data rate and high transmission
analyser. It is noted that energy consumption of the same power for the end-devices far away from the gateway. On
packet transmission varies by more than 50% between the studying the effect of environmental factors, it is shown that
maximum and minimum values. This measurement stresses signal strength is decreased by 6dBm at 60◦ C. Even this small
on the usage of ADR to reduce energy consumption. deviation can increase packet loss in the messages transmitted
Angrisani et al. [40] assess the performance of LoRa by end-devices far away from the gateway.
under critical noise conditions. Transmitter, powered through Mikhaylov et al. [67] study LoRaWAN’s susceptibility to
a power bank, placed 10m away from the receiver. A White inter-network interference. Experiments with and without an
Gaussian noise is generated to corrupt the transmitted signal. interferer between transmitter and receiver gives an insight
The fixed parameters distance, payload and preamble are set to design a protocol for finding dynamic communication
to 10m, 1 byte and 8 symbols respectively. On varying SF, parameters. Experimental result shows that a personalized
BW and CR to all possibilities, authors claim that an increase communication parameter for each end-device will aid scala-
in BW increases the packet loss with lower SFs. But packet bility.
loss is decreased with larger SFs. The authors also state that Research Solutions. Fair Adaptive Datarate Algorithm
an increase in CR can trade off an increase in BW and SF. (FADR) [62] is proposed to select SFs and transmission power
Finally, it is concluded that LoRa is highly robust to high to achieve data extraction rate among all end-devices. SF is
noise levels and recommend further investigation by varying allocated based on the method described in [63], using RSSI
the fixed parameters of this experiment. Iova et al. [112] and power levels. End-devices are grouped based on regions.
investigate the performance of LoRa in mountain regions This technique simulated in LoRaSim achieves 300% higher
and identify the factors affecting transmission parameters. fairness than the technique proposed by Bor et al. [21] and
Hakkenberg et al. [113] and Neumann et al. [114] evaluate the 22% higher fairness than the technique proposed by Reynders
performance of LoRa in both indoor and outdoor environments et al. [63] while reducing network energy consumption by
and recommend that transmission parameters have to be varied 22%.
according to the deployed environment. Bor et al. [21] consider bandwidth and transmission power
Describing the operations of LoRaWAN, Augustin et al. for scalability analysis through simulation. Georgiou et al. [57]
[115] evaluate LoRa’s receiver sensitivity, network coverage develop a mathematical model for a LoRaWAN network with
using Freescale KRDM-KL25Z development board [116] with a single gateway by also considering other unique LoRaWAN
Semtech 1276 transceiver [95] as end-device and Cisco 910 features like modulation and radio duty-cycling. A mathemati-
industrial router [117] as gateways. Gateway is connected cal investigation on link outage, considering signal below SNR
to The Things Network server to monitor received packets. threshold and capture effect, is carried out in order to study
Gateway is placed indoors and end-devices were kept moving their effects on scalability. It is inferred that the latter reduces
outdoors in the urban environment. Transmit power of the network performance with increase in density of end-devices,
end-devices was set to minimum 2dBm with a 3-dBi antenna. which hinders the scalability.
Packet losses start at 100m. The measured RSSI values were Different from [21], which studies the scalability of Lo-
slightly above the specified values for each SF. For network RaWAN network through LoRaSim, Mikhaylov et al. [37]
coverage experiments, the gateway was placed 5 m above present mathematical analysis without considering many fac-
the ground level and end-devices were kept in a car with tors. Bor et al. [21] build a simulator ruminating Bandwidth
default transmit power 14dBm specified in [118]. PRR is and Transmission Power for modeling uplink behavior. Three
tested for SF7, SF9, SF12 at various distances with ACK and experiments are conducted. The first experiment with a single
retransmission turned off. At 2800m, SF7 achieved 0% PRR gateway and multiple end-devices with homogeneous commu-
while SF12 delivers about 80% of the packet. The authors nication parameter infers that a single gateway can support
find that communication coverage is directly proportional to 120 end-devices per 3.8 hectares. The second experiment
SF values. contains a single gateway and heterogeneous communication
Blenn et al. [41] analyse the 9.4GB data obtained from parameters, such that the end-device’s uplink air time is
the The Things Network during December 2015 and July decreased, showing 13× increase in node density than the pre-
2016 from 1618 unique devices. It is inferred that 3.7% of vious experiment. The final experiment with multiple gateways
unique packets were received by two gateways, 1.1% of unique improves data extraction rate. Two suggested guidelines are
packets were received by three gateways. Average payload size to develop a protocol to decide communication parameters
is 18 Bytes where 93.7% of captured payloads are less than dynamically and to evaluate optimal gateway placement for
50 bytes and 50% of the payloads are less than 19 bytes. It is better scalability.
observed that using higher SF and higher transmission power Chall et al. [79] collect empirical data in Lebanon to verify
results in low packet loss. various radio propagation models like Okumura-Hata [119],
 13

[120], Cost-231 Hata [121] to find their drawbacks and fix the transmission power are probed. If a potential setting is not
them with additional proposals to make it acceptable for LoRa found, the algorithm employs an iteration bound to try other
networks. Bor et al. [21] also propose a mathematical model settings. It is shown that the proposed probing algorithm finds
for LoRa communication coverage based on the empirical an optimal setting that uses only 44% more energy than the
data obtained over 2.6 Km of rural area and 100 m of built- ideal setting within 285 probes.
up environment. Demetri et al. [78] compare link attenuation Van et al. [66] conduct experiment with single, multiple
of LoRa signals in free space and Bor’s model. It is shown gateways and various SF to study the scalability of LoRaWAN
that free-space model underestimates signal attenuation while in NS-3 simulator. Error model combined with NS-3 Lo-
Bor’s model overestimates it. It is also claimed that Bor’s RaWAN protocol is constructed through extensive baseband
model [21] need on-site measurements which is hard due bit error rate simulations and used as an interference model.
to that most of the covered regions comprise of transitional Experimental results show that usage of ACKs severely affects
links, which are defined as links with dynamic temporal link uplink traffic and having multiple gateways improves scalabil-
qualities. Hence, a new automated link quality estimation ity to a smaller extent. It is also showed that assigning dynamic
system without requiring on-site measurements is developed communication parameters will help to up-scale node density.
by Demetri et al. [78]. To achieve this, remote sensing spectral Reynders et al. [63] find an optimal SF setting to reduce
images from an open-source satellite is used. These images collision probability and distribute SFs and transmission pow-
are fed as input to Support Vector Machine (SVM) [122] ers to decrease Packet Error Ratio (PER) of end-devices far
to classify different constitution of land coverage like water away from the gateway. A routine to assign SFs and power
bodies, forests and buildings. Okumura-Hata model [119], control is developed based on genetic algorithm. The key idea
[120] is modified to support LoRa link estimation on different of this algorithm is to assign different SFs and power control
land coverage. The results of SVM classification is used to different nodes such that signals do not interfere with each
to automatically choose and configure parameters of link other. Simulation of this technique in NS-3 shows that the PER
estimation model. of the overall network is reduced by 42% and the packet error
Margin et al. [26] implement a new NS-3 module to ratio of end-devices far away from the gateway is reduced by
simulate dense urban environments. The link performance 50%.
and measurements, signal attenuation due to buildings and Pop et. al [64] extends LoRaSim [21] by adding more
other factors are given. Spreading Factor assignment is done features like ACK, downlink data messages and presents a new
based on power levels of the end-device at the gateway. The simulator called LoRaWANSim. Same experimental settings
gateway will bind with the end-device transmitting on highest used in [21] are used with additional downlink traffic to
received power level. 17 Gateways are placed in a hexagonal study the scalability with downlink ACKs. It is inferred that
grid around the central gateway, covering a 7.5 Km radius. scalability is hampered as handling ACKs reduce network
Totally 104 end-devices are placed randomly. The experimental performance.
results show that densifying gateways such that each gateway Voigt et al. [69] compare the usage of directional antenna
covers 1200m can achieve a packet delivery over 90%. But and multiple gateways to alleviate interference that arises due
this increases collision as the number of end-devices using to dense deployments. It is shown that the gain of multiple
SF7 increases. The authors recommend that ADR mechanisms gateways outperforms the usage of directional antennae.
should be leveraged to counter such collisions. Key Insights. The key insights regarding error correction
Two techniques are proposed by Cumo et. al [65] to allocate are listed as follows.
SFs to end-devices. First technique EXPLoRa-SF allocates SFs • Communications with higher SF and higher TP can reach
based on RSSI of the end-device received by the gateway. a longer distance [41], [115].
EXPLoRa-AT guarantees Time-on-Air equalization for all • Network density can be scaled-up by adding more gate-
end-devices in addition to SF allocation. This is achieved by ways and personalizing transmission parameters like data
ordered waterfilling technique to evenly distribute channel load rate and channel allocation using ADR for each end-
among end-devices in the network. Simulation of EXPLoRa- device [25], [66], [67], [69].
SF and EXPLoRa-AT in ”LoRaSim” performs better than the • Energy consumption varies up to 50% while using lowest
basic ADR. and highest transmission power to transmit the same
Bor et al. [123] study the impact of transmission param- packet. This shows that the lowest possible transmission
eter selection on communication performance and propose power should be used for saving maximum energy.
an algorithm to quickly identify the optimal transmission • It is not always required to increase SF if some locations
parameters for energy efficiency and reliable communication. cannot reach gateway. Increasing bandwidth on the same
It is shown that investing high energy for higher SF values SF also increases PRR [34]. But this is not always the
does not always improve communication performance. From case at lower SF [40]. Investing more energy by using
experimental results, authors claim that it is also possible to high SF does not necessarily improve communication
achieve minimum energy efficiency while selecting desired performance [123]. Hence, varying transmission parame-
transmission parameters based on application requirements. ters based on indoor/outdoor deployed environment gives
The proposed probing algorithm finds a transmission param- better performance [113], [114].
eter that halves the transmission power with PRR larger than • Only 10% difference is identified in PRR between fastest
a threshold. If not found, other settings that uses at most half and lowest SF and TP settings for the end-devices located
 14

at the farthest reachable point from gateways. Hence, message. DevNonce should be unique for each Join Request.
lower SF and TP settings can help reduce resource Network Server declines or excludes the node sending Join-
consumption [68]. request with a previously used DevNonce. This paper identifies
• Besides, the communication performance can also be the probability of regenerating a used DevNonce and the
improved by reducing unconfirmed messages [63], [66]. scenario where malicious node floods Network Server to
register possible DevNonce’s randomly, making the future join
request of non-malicious nodes tougher. A random number
E. LoRa Security
generator algorithm is proposed and the size of DevNonce is
Measurements. On describing the LoRa network stack, increased from 16 to 24 bits to overcome these shortcomings.
Arsas et al. [47] explore the vulnerabilities of LoRa. This Kim et al. [71] focus on resolving the following three
paper expounds four possible techniques to compromise the problems. The first problem is the current DevNonce system.
LoRa network. Firstly, compromising security keys. This is Join-request sent by benign end-device can be mistaken as a
easier if an attacker can gain physical access to an end- replay attack by network server because an end-device can
device. Feasibility of this attack is demonstrated through regenerate old DevNonce. Not storing all the past DevNonce’s
experiments. Extracting security keys from any end-device will will not prevent replay attack. The second problem is the 24-
enable the attacker to decrypt any message in the network. bit DevNonce proposed in the article [76], that is incompatible
Secondly, The Jamming attack. The Jamming attack is an with the current LoRaWAN specification. The third problem is
attack in which the communication channel is jammed with an the token-based scheme proposed in [72]. This prevents replay
intentional interference to corrupt the data signal sent in that attack effectively but does not consider the scenario where the
channel. Reynders et al. [124] show that LoRa is also prone token is lost. For example, when the end-device reboots, the
to Jamming attacks even if the chirp spread spectrum is robust token is lost. This paper proposes two types of Join-requests
to interferences. Demonstrating this through experiments, it is called Initial and Non-Initial Join requests. Non-Initial Join-
shown that a specific node can also be targeted. Thirdly, The request uses the token and it changes after each join procedure
replay attack. It is an attack in which the intruder intercepts the is complete. Current DevNonce may be regenerated. As the Ini-
message and resends it to the receiver whenever the intruder tial Join-request rarely occurs, regeneration of old DevNonce
wants to accomplish a particular task. Miller et al. [49] show is negligible. When the token is lost, the node reboots and
that the consequence of this attack depends on the application initiates Initial Join-request. These techniques are proved to
scenarios. Finally, the wormhole attack captures a packet from enhance security through theory and experiments.
a non-malicious node and this never reaches the server. Some Oniga et. al [50] explicates different security aspects of
credentials stored in the packet is valid and can be used at any LoRa and proposes a secure network architecture framework.
time in the network. Implementation of this model under different testing scenarios
Butun et al. [48] surveys and verifies the feasible security recommend techniques for better data security and privacy of
threats of LoRa V1.1 with Scyther security verification tool LoRa based applications.
[125]. The attacks verified are (i) RF Jamming attack, (ii) Third-party authorisation. Girard [73] pointed out that both
Replay attack, (iii) Class B beacon synchronization attack, (iv) application and network session keys being generated at the
Network traffic analysis, (v) Man-In-The-Middle attack. network server will create a conflict of interest between
Donmez et al. [58] identify the security vulnerabilities of network and application service providers. The network server
V1.1 in the backward compatibility scenario. The security and application server can derive both Application and Net-
vulnerabilities of LoRaWAN V1.0.2 and corresponding so- work session keys which is not secure if two different organi-
lutions added in LoRaWAN specification V1.1 to mitigate zations are involved. So, a trusted third-party key management
them are discussed. The open vulnerabilities during backward architecture is proposed.
compatibility scenarios are discussed and countermeasures are The problem of Key management and update mechanism is
proposed to mitigate them. While the specification is explain- well addressed by the techniques proposed in [129], [130] for
ing only one backward compatibility scenario, this article wireless sensor networks. As Girard [73] introduces a third-
verifies all possible scenarios to find other vulnerabilities. party architecture, communication overhead is increased that
Research Solutions. The discussion on proxy-based key may degrade the network performance. So, Kim et al. [74]
establishment for securing messages in the IoT context in propose a dual key based activation scheme to support key
[126], [127], [128], gives an insight on the proxy-based generation and update without adding any complexities. This
key establishment for LoRa. Naoui et al. [75] discuss the paper explicates that AppKey which is not updated periodically
possibilities of applying proxy-based key exchange systems will pose many security threats through which an attacker
to enhance LoRa security. Bit flipping attack, an attack by can steal all transmitted data of a target node. The proposed
which the bits of ciphertext is changed, is countered in [70] technique of this paper separates the Network and Application
using circular shift and swap techniques. Session key generation to appropriate servers. These keys
Key Generation. Tomasin et al. [76] analyse the security cannot be derived from a public key and not shared with other
of join procedures, especially through On-The-Air-Activation devices. This scheme is proven to be both delay and power
(OTAA) of LoRa. Join procedures are activated at least once- feasible through experiments.
a-day to check whether the node is still connected to the Trust mechanism and blockchains. Some works [131]–
Network. DevNonce is a random number in the Join-request [136], discuss the application of blockchain technology for
 15

IoT scenarios. This is helpful to apply blockchain to LoRa larger than the retransmission time. This explains the need for
networks. Lin et al. [137] build a trust mechanism for LoRa dynamic retransmission policies. A joint retransmission policy
using blockchain technology as attacking a blockchain system with channel coding and compression is theoretically studied,
is computationally difficult as the attacker has to transcend but the implementation of such computationally expensive
at least half of the system’s computational ability. The pro- techniques on constrained LoRa hardware must be considered
posed system implements Blockchain manager component to in future. Besides, it has to be noted that a modular retransmis-
network server. This framework is proposed for large scale sion policy without any dependencies on other techniques is
deployments of LoRa like wild-life monitoring, asset tracking inevitable. None of the techniques has addressed this problem
and smart parking. of varying retransmission timers dynamically. The factors
Key Insights. The key insights regarding the security issues triggering retransmission, even in the case of correct reception,
are listed as follows. must be extensively studied through experiments for various
• An attacker can decrypt messages in the network by scenarios and a dynamic policy must be devised to improve
compromising security keys if they can get physical the performance of LoRa.
access to end-devices.
• LoRa devices are susceptible to jamming attack, replay C. Communication Range
attack, beacon synchronization attack, traffic analysis and
man-in-the-middle attack. Performance evaluations in [32]–[37] expound the need to
• Larger size DevNonce prevents join attacks [76].
provide reliable transmissions for long range LoRa links. Some
• Application and Network key generation and update has
of the techniques like Du et al. [96] improves communication
to be separated based on the application scenario. range for wireless sensor networks without heavy hardware
• Blockchain mechanisms can also be implemented on Low
modifications. The communication range could be further
powered devices [137]. enlarged in the future as LoRa chips will evolve to support
new functionalities.
VI. O PEN I SSUES Choir [51] improves the range but its implementation on
commercial radio chips is ambiguous as they were experi-
As introduced in V, various techniques have been proposed
mented on USRP radios and may require modifying the com-
to address the challenges of LoRa deployments. Some so-
mercial radio. NetScatter [55] theoretically proves that Choir
lutions still leave room to further improve the performance
[51] cannot scale-up well. NetScatter [55] leverages backscat-
of LoRa. For example, some solutions for choosing dynamic
ter with distributed coding to enable concurrent transmission
communication parameters consider most of the factors, but
of 256 nodes which still needs some hardware modifications
did not take into account the ambient temperature which plays
on commercial chips. Even though the above techniques
a major role in reducing the signal strength [68]. Based on the
Choir [51] and NetScatter [55] have improved the range,
above analysis of research challenges and recently proposed
communication range still needs to be improved without heavy
solutions in section IV and V, some open issues of the LoRa
modifications of the commercially available chips.
technology are presented in this section.

A. Optimal placement of multiple gateways: D. Security

Some works, [38], [21], [66], [69] aiding scalability and Security spans over a range of attacks like node imper-
interference use multiple gateways as a solution. Even though sonation, eavesdropping, Black hole attack, Wormhole attack,
these techniques outperform existing results, using multiple etc., as discussed by Zhou et al. [86]. Only few key man-
gateways instigate to study the optimal gateway placements agement techniques have been discussed and proposed for
for LoRa deployments. Optimal placement of gateways is LoRa. Several attacks still need to be addressed to secure
always dependent on the application and constraints of the LoRa networks. Even though some techniques enhance the
hardware used in the application. A generic solution for security of existing LoRa standard, security requirements of
optimal placement of gateways according to the categories of LoRa are not discussed based on the applications of LoRa. In
applications will further improve the performance. the future, each application of LoRa will demand their own
security needs. For example, some applications may require
B. Link Co-ordination: Network and Application Session keys to be independent and
Network Session Key should be confidential from application
Countering degradation due to downlink ACKs: Some
server and vice-versa. Hence, the security needs of each
works like [39] and [64] state that downlink ACKs reduce
deploying scenario have to be deeply investigated to mitigate
network performance, as end devices are not able to transmit
vulnerabilities arising due to different application scenarios.
subsequent packets if downlink ACKs are delayed or cor-
rupted. This gives rise to the need for developing a dynamic
ACK mechanism to improve network performance. VII. C ONCLUSION
Dynamic retransmission policies: One solution to counter Among different LPWAN technologies, LoRa networking
downlink ACKs will be setting dynamic retransmission poli- is widely adopted, since it allows to build and maintain an
cies. Static retransmission policy degrades the network perfor- autonomous network without third-party infrastructure, while
mance when the time taken by an ACK to reach the end node is satisfying the low power and long range communication
 16

requirements. By investigating the challenges faced during tering applications,” Sep. 2017, white paper.
deploying LoRa networks, recent solutions developed are dis- [26] D. Magrin, M. Centenaro, and L. Vangelista, “Performance evaluation
of LoRa networks in a smart city scenario,” in Proceedings of the
cussed in detail. Based on the challenges and recent solutions, International Conference On Communications, 2017, pp. 1–7.
we present some open issues that need to be addressed for [27] W. Du, Z. Xing, M. Li, B. He, L. H. C. Chua, and H. Miao, “Optimal
practical large-scale deployment of LoRa networks. sensor placement and measurement of wind for water quality studies
in urban reservoirs,” in Proceedings of the International symposium on
Information processing in sensor networks, 2014, pp. 167–178.
R EFERENCES [28] W. Du, Z. Xing, M. Li, B. He, L. H. C. Chua, and H. Miao, “Sensor
placement and measurement of wind for water quality studies in urban
[1] J. Lin, W. Yu, N. Zhang, X. Yang, H. Zhang, and W. Zhao, “A survey
reservoirs,” ACM Transactions on Sensor Networks, vol. 11, no. 3, pp.
on internet of things: Architecture, enabling technologies, security and
41:1–41:27, 2015.
privacy, and applications,” IEEE Internet of Things Journal, vol. 4,
[29] W. Du, J. C. Liando, H. Zhang, and M. Li, “Pando: Fountain-enabled
no. 5, pp. 1125–1142, 2017.
fast data dissemination with constructive interference,” IEEE/ACM
[2] A. Al-Fuqaha, M. Guizani, M. Mohammadi, M. Aledhari, and
Transactions on Networking, vol. 25, no. 2, pp. 820–833, 2017.
M. Ayyash, “Internet of things: A survey on enabling technologies,
[30] W. Du, J. Liando, H. Zhang, and M. Li, “When pipelines meet fountain:
protocols, and applications,” IEEE Communications Surveys & Tutori-
Fast data dissemination in wireless sensor networks,” in Proceedings of
als, vol. 17, no. 4, pp. 2347–2376, 2015.
the ACM Conference on Embedded Networked Sensor Systems, 2015,
[3] J. A. Stankovic, “Research directions for the internet of things,” IEEE
pp. 365–378.
Internet of Things Journal, vol. 1, no. 1, pp. 3–9, 2014.
[31] J.Catalano, J-P.Coupigny, M.Kuyper, N.sornin, and A.Yegin, “Fuota
[4] “5G americas LTE progress leading to the 5G massive Internet of
process summary technical recommendation,” 2019.
Things,” 2017.
[32] R. Fuidiak, P. Mlynek, J. Misurec, and M. Strajt, “Simulated coverage
[5] “Northstream white paper, Massive IoT: Different technologies for
estimation of single gateway LoRaWAN network,” in Proceedings of
different needs,” 2017.
the International Conference on Systems, Signals and Image Process-
[6] Ericsson, “Ericsson mobility report,” 2017.
ing, 2018, pp. 1–4.
[7] V. Gazis, “A survey of standards for machine-to-machine and the
[33] J. Petajajarvi, K. Mikhaylov, A. Roivainen, T. Hanninen, and M. Pet-
internet of things,” IEEE Communications Surveys & Tutorials, vol. 19,
tissalo, “On the coverage of LPWANs: range evaluation and channel
no. 1, pp. 482–511, 2017.
attenuation model for LoRa technology,” in Proceedings of the Inter-
[8] S. Chen, R. Ma, H.-H. Chen, H. Zhang, W. Meng, and J. Liu,
national Conference on ITS Telecommunications, 2015, pp. 55–59.
“Machine-to-machine communications in ultra-dense networksa sur-
[34] J. Petäjäjärvi, K. Mikhaylov, R. Yasmin, M. Hämäläinen, and J. Iinatti,
vey,” IEEE Communications Surveys & Tutorials, vol. 19, no. 3, pp.
“Evaluation of LoRa LPWAN technology for indoor remote health and
1478–1503, 2017.
wellbeing monitoring,” International Journal of Wireless Information
[9] U. Raza, P. Kulkarni, and M. Sooriyabandara, “Low power wide area
Networks, vol. 24, no. 2, pp. 153–165, 2017.
networks: An overview,” IEEE Communications Surveys & Tutorials,
[35] N. Harris and J. Curry, “Development and range testing of a LoRaWAN
vol. 19, no. 2, pp. 855–873, 2017.
system in an urban environment,” World Academy of Science, Engi-
[10] H. Wang and A. O. Fapojuwo, “A survey of enabling technologies of
neering and Technology, International Journal of Electrical, Computer,
low power and long range machine-to-machine communications,” IEEE
Energetic, Electronic and Communication Engineering, vol. 12, no. 1,
Communications Surveys & Tutorials, vol. 19, no. 4, pp. 2621–2639,
pp. 43–51, 2018.
2017.
[36] J. Haxhibeqiri, A. Karaağaç, F. Van den Abeele, W. Joseph, I. Moer-
[11] D. Flore, “3GPP standards for the Internet-of-Things,” Recuperado el,
man, and J. Hoebeke, “LoRa indoor coverage and performance in an
vol. 25, 2016.
industrial environment: Case study,” in Proceedings of the International
[12] J. C. Zuniga and B. Ponsard, “Sigfox system description,” LPWAN@
Conference on Emerging Technologies And Factory Automation, 2017,
IETF97, Nov. 14th, 2016.
pp. 1–8.
[13] L. Vangelista, A. Zanella, and M. Zorzi, “Long-range IoT technologies:
[37] K. Mikhaylov, J. Petaejaejaervi, and T. Haenninen, “Analysis of
The dawn of LoRa,” in Proceedings of the Future Access Enablers of
capacity and scalability of the LoRa low power wide area network
Ubiquitous and Intelligent Infrastructures, 2015, pp. 51–58.
technology,” in Proceedings of the European Wireless Conference,
[14] J. Lee, Y. Kim, Y. Kwak, J. Zhang, A. Papasakellariou, T. Novlan,
2016, pp. 1–6.
C. Sun, and Y. Li, “LTE-advanced in 3GPP rel-13/14: An evolution
[38] Semtech, “LoRaWAN’s capacity trials in dense urban environments.”
toward 5G,” IEEE Communications Magazine, vol. 54, no. 3, pp. 36–
[39] D. Bankov, E. Khorov, and A. Lyakhov, “On the limits of LoRaWAN
42, 2016.
channel access,” in Proceedings of the International Conference on
[15] M. Mezzavilla, M. Zhang, M. Polese, R. Ford, S. Dutta, S. Rangan,
Engineering and Telecommunication, 2016, pp. 10–14.
and M. Zorzi, “End-to-end simulation of 5G mmwave networks,” IEEE
[40] L. Angrisani, P. Arpaia, F. Bonavolontà, M. Conti, and A. Liccardo,
Communications Surveys & Tutorials, vol. 20, no. 3, pp. 2237–2263,
“LoRa protocol performance assessment in critical noise conditions,” in
2018.
Proceedings of the International Forum on Research and Technologies
[16] R. S. Sinha, Y. Wei, and S.-H. Hwang, “A survey on LPWA technology:
for Society and Industry, 2017, pp. 1–5.
LoRa and NB-IoT,” ICT Express, vol. 3, no. 1, pp. 14–21, 2017.
[41] N. Blenn and F. Kuipers, “LoRaWAN in the wild: Measurements from
[17] Ingenu, “How RPMA works?” White Paper, 2016.
the things network,” arXiv preprint arXiv:1706.03086, 2017.
[18] W. Webb, Understanding Weightless: Technology, Equipment, and
[42] M. Lauridsen, B. Vejlgaard, I. Z. Kovacs, H. Nguyen, and P. Mogensen,
Network Deployment for M2M Communications in White Space. Cam-
“Interference measurements in the European 868 MHz ISM band
bridge University Press, 2012.
with focus on LoRa and SigFox,” in Proceedings of the Wireless
[19] LoRa-Alliance, “A technical overview of LoRaWAN,” Technical Mar-
Communications and Networking Conference, 2017, pp. 1–6.
keting Workgroup, Nov. 2015.
[43] B. Vejlgaard, M. Lauridsen, H. Nguyen, I. Z. Kovács, P. Mogensen, and
[20] R. Fujdiak, P. Blazek, K. Mikhaylov, L. Malina, P. Mlynek, J. Misurec,
M. Sorensen, “Coverage and capacity analysis of Sigfox, LoRa, GPRS,
and V. Blazek, “On track of sigfox confidentiality with end-to-end
and NB-IoT,” in Proceedings of the Vehicular Technology Conference,
encryption,” in Proceedings of the 13th ACM International Conference
2017, pp. 4–7.
on Availability, Reliability and Security, 2018, p. 19.
[44] G. Zhu, C.-H. Liao, M. Suzuki, Y. Narusue, and H. Morikawa,
[21] M. C. Bor, U. Roedig, T. Voigt, and J. M. Alonso, “Do LoRa low-
“Evaluation of LoRa receiver performance under co-technology inter-
power wide-area networks scale?” in Proceedings of the International
ference,” in Proceedings of the Annual Consumer Communications &
Conference on Modeling, Analysis and Simulation of Wireless and
Networking Conference, 2018, pp. 1–7.
Mobile Systems, 2016, pp. 59–67.
[45] C. Orfanidis, L. M. Feeney, M. Jacobsson, and P. Gunningberg, “Inves-
[22] Semtech, “Smart buildings transformed using semtech’s LoRa technol-
tigating interference between LoRa and IEEE 802.15.4g networks,” in
ogy,” Apr. 2017, white paper.
Proceedings of the Wireless and Mobile Computing, Networking and
[23] Semtech, “Smart cities transformed using semtech’s LoRa technology,”
Communications, 2017, pp. 1–8.
Jul. 2017, white paper.
[46] G. Ferré, “Collision and packet loss analysis in a LoRaWAN network,”
[24] Semtech, “Revolutionising smart agriculture using semtech’s LoRa
in Proceedings of the European Signal Processing Conference, 2017,
technology,” Oct. 2017, white paper.
pp. 2586–2590.
[25] Semtech, “Real-world LoRaWAN network capacity for electrical me-
 17

[47] E. Aras, G. S. Ramachandran, P. Lawrence, and D. Hughes, “Exploring [70] J. Lee, D. Hwang, J. Park, and K.-H. Kim, “Risk analysis and
the security vulnerabilities of LoRa,” in Proceedings of the Interna- countermeasure for bit-flipping attack in LoRaWAN,” in Proceedings
tional Conference on Cybernetics, 2017, pp. 1–6. of the International Conference on Information Networking, 2017, pp.
[48] I. Butun, N. Pereira, and M. Gidlund, “Analysis of LoRaWAN v1. 1 549–551.
security,” in Proceedings of the MobiHoc Workshop on Experiences [71] J. Kim and J. Song, “A simple and efficient replay attack prevention
with the Design and Implementation of Smart Objects, 2018, pp. 5–10. scheme for LoRaWAN,” in Proceedings of the International Conference
[49] R. Miller, “Lora security: Building a secure lora solution,” MWR Labs on Communication and Network Security, 2017, pp. 32–36.
Whitepaper, 2016. [72] S. Na, D. Hwang, W. Shin, and K.-H. Kim, “Scenario and countermea-
[50] B. Oniga, V. Dadarlat, E. De Poorter, and A. Munteanu, “Analysis, sure for replay attack using join request messages in LoRaWAN,” in
design and implementation of secure LoRaWAN sensor networks,” in Proceedings of the International Conference on Information Network-
Proceedings of the International Conference on Intelligent Computer ing, 2017, pp. 718–720.
Communication and Processing, 2017, pp. 421–428. [73] P. Girard, “Low power wide area network security,” gematlo, Tech.
[51] R. Eletreby, D. Zhang, S. Kumar, and O. Yağan, “Empowering low- Rep., Dec. 2015.
power wide area networks in urban settings,” in Proceedings of the [74] J. Kim and J. Song, “A dual key-based activation scheme for secure
Conference of the Special Interest Group on Data Communication, LoRaWAN,” Wireless Communications and Mobile Computing, vol.
2017, pp. 309–321. 2017, pp. 1–12, 2017.
[52] A. Dongare, R. Narayanan, A. Gadre, A. Luong, A. Balanuta, S. Ku- [75] S. Naoui, M. E. Elhdhili, and L. A. Saidane, “Enhancing the security of
mar, B. Iannucci, and A. Rowe, “Charm: Exploiting geographical diver- the IoT LoRaWAN architecture,” in Proceedings of the International
sity through coherent combining in low-power wide-area networks,” in Conference on Performance Evaluation and Modeling in Wired and
Proceedings of the International Conference on Information Processing Wireless Networks, 2016, pp. 1–7.
in Sensor Networks, 2018, pp. 60–71. [76] S. Tomasin, S. Zulian, and L. Vangelista, “Security analysis of Lo-
[53] Y. Peng, L. Shangguan, Y. Hu, Y. Qian, X. Lin, X. Chen, D. Fang, RaWAN join procedure for internet of things networks,” in Proceedings
and K. Jamieson, “PLoRa: a passive long-range data network from of the Wireless Communications and Networking Conference Work-
ambient lora transmissions,” in Proceedings of the Conference of the shops, 2017, pp. 1–6.
Special Interest Group on Data Communication, 2018, pp. 147–160. [77] J. Haxhibeqiri, E. De Poorter, I. Moerman, and J. Hoebeke, “A survey
[54] V. Talla, M. Hessar, B. Kellogg, A. Najafi, J. R. Smith, and S. Gol- of lorawan for iot: From technology to application,” Sensors, vol. 18,
lakota, “LoRa backscatter: Enabling the vision of ubiquitous con- no. 11, p. 3995, 2018.
nectivity,” Proceedings of the on Interactive, Mobile, Wearable and [78] S. Demetri, M. Zúñiga, G. P. Picco, F. Kuipers, L. Bruzzone, and
Ubiquitous Technologies, vol. 1, no. 3, p. 105, 2017. T. Telkamp, “Automated estimation of link quality for lora: A remote
[55] M. Hessar, A. Najafi, and S. Gollakota, “Netscatter: Enabling large- sensing approach,” in Proceedings of the International Conference on
scale backscatter networks,” arXiv preprint arXiv:1808.05195, 2018. Information Processing in Sensor Networks (IPSN), 2019.
[56] P. J. Marcelis, V. S. Rao, and R. V. Prasad, “DaRe: Data recovery [79] R. El Chall, S. Lahoud, and M. El Helou, “Loraicdcsnetwork: Radio
through application layer coding for LoRaWAN,” in Proceedings of propagation models and performance evaluation in various environ-
the International Conference on Internet-of-Things Design and Imple- ments in lebanon,” IEEE Internet of Things Journal, 2019.
mentation, 2017, pp. 97–108. [80] J. C. Liando, A. Gamage, A. W. Tengourtius, and M. Li, “Known and
[57] O. Georgiou and U. Raza, “Low power wide area network analysis: unknown facts of lora: Experiences from a large-scale measurement
Can LoRa scale?” IEEE Wireless Communications Letters, vol. 6, no. 2, study,” ACM Transactions on Sensor Networks (TOSN), vol. 15, no. 2,
pp. 162–165, 2017. p. 16, 2019.
[58] T. C. Dönmez and E. Nigussie, “Security of LoRaWAN v1. 1 in [81] Semtech, “Smart parking,” https://www.semtech.com/lora/
backward compatibility scenarios,” Procedia Computer Science, vol. lora-applications, White Paper.
134, pp. 51–58, 2018. [82] Semtech, “Smart lighting,” https://www.semtech.com/lora/
[59] M. Sandell and U. Raza, “Application layer coding for iot: benefits, lim- lora-applications, White Paper.
itations, and implementation aspects,” IEEE Systems Journal, no. 99, [83] Semtech, “Lora technology: Transforming golf courses with iot,” Tech.
pp. 1–8, 2018. Rep., 2018.
[60] J. Haxhibeqiri, F. Van den Abeele, I. Moerman, and J. Hoebeke, “LoRa [84] Semtech, “Lora technology: Mallorca develops first lorawan smart
scalability: A simulation model based on interference measurements,” island,” Tech. Rep., 2018.
Sensors, vol. 17, no. 6, p. 1193, 2017. [85] Z. Qin and J. A. McCann, “Resource efficiency in Low-Power Wide-
[61] B. Reynders, Q. Wang, P. Tuset-Peiro, X. Vilajosana, and S. Pollin, Area Networks for IoT applications,” in Proceedings of the Global
“Improving reliability and scalability of LoRaWANs through Communications Conference, 2017, pp. 1–7.
lightweight scheduling,” IEEE Internet of Things Journal, 2018. [86] Y. Zhou, Y. Fang, and Y. Zhang, “Securing wireless sensor networks:
[62] K. Q. Abdelfadeel, V. Cionca, and D. Pesch, “A fair adaptive data rate a survey,” IEEE Communications Surveys & Tutorials, vol. 10, no. 3,
algorithm for LoRaWAN,” arXiv preprint arXiv:1801.00522, 2018. 2008.
[63] B. Reynders, W. Meert, and S. Pollin, “Power and spreading factor [87] F. Orfei, C. B. Mezzetti, and F. Cottone, “Vibrations powered LoRa
control in low power wide area networks,” in Proceedings of the sensor: An electromechanical energy harvester working on a real
International Conference on Communications, 2017, pp. 1–6. bridge,” in Proceedings of the Sensors, 2016, pp. 1–3.
[64] A.-I. Pop, U. Raza, P. Kulkarni, and M. Sooriyabandara, “Does [88] A. Varshney, O. Harms, C. Pérez-Penichet, C. Rohner, F. Hermans,
bidirectional traffic do more harm than good in LoRaWAN based and T. Voigt, “Lorea: A backscatter architecture that achieves a long
LPWA networks?” arXiv preprint arXiv:1704.04174, 2017. communication range,” in Proceedings of the ACM Conference on
[65] F. Cuomo, M. Campo, A. Caponi, G. Bianchi, G. Rossini, and Embedded Network Sensor Systems, 2017, p. 18.
P. Pisani, “EXPLoRa: Extending the performance of LoRa by suitable [89] J. Navarro-Ortiz, S. Sendra, P. Ameigeiras, and J. M. Lopez-Soler,
spreading factor allocations,” in Proceedings of the Wireless and Mobile “Integration of LoRaWAN and 4G/5G for the Industrial Internet of
Computing, Networking and Communications, 2017, pp. 1–8. Things,” IEEE Communications Magazine, vol. 56, no. 2, pp. 60–67,
[66] F. Van den Abeele, J. Haxhibeqiri, I. Moerman, and J. Hoebeke, 2018.
“Scalability analysis of large-scale LoRaWAN networks in NS-3,” [90] Ideetron, “LORANK gateway, ideetron lora products,” url =
Internet of Things Journal, vol. 4, no. 6, pp. 2186–2198, 2017. https://webshop.ideetron.nl/LoRa.
[67] K. Mikhaylov, J. Petäjäjärvi, and J. Janhunen, “On LoRaWAN scala- [91] W. Gao and W. Du and Z. Zhao and G. Min and M. Singhal, “Towards
bility: Empirical evaluation of susceptibility to inter-network interfer- Energy-Fairness in LoRa Networks,” in IEEE ICDCS, 2019.
ence,” in Proceedings of the European Conference on Networks and [92] P. Hu, P. Zhang, M. Rostami, and D. Ganesan, “Braidio: An integrated
Communications, 2017, pp. 1–6. active-passive radio for mobile devices with asymmetric energy bud-
[68] M. Cattani, C. A. Boano, and K. Römer, “An experimental evaluation of gets,” in Proceedings of the 2016 ACM SIGCOMM Conference. ACM,
the reliability of LoRa long-range low-power wireless communication,” 2016, pp. 384–397.
Journal of Sensor and Actuator Networks, vol. 6, no. 2, p. 7, 2017. [93] WIMOD, “WIMOD - IM880A Data sheet,” WiMOD iM880A , Tech.
[69] T. Voigt, M. Bor, and U. Roedig, “Mitigating inter-network interference Rep., Jul. 2017.
in LoRa networks,” in Proceedings of the MadCom: New Wireless [94] Kerlink, “Wirnet station datasheet,” Tech. Rep., Jan. 2016.
Communication Paradigms for the Internet of Things, 2017, pp. 323– [95] Semtech, “Sx1272/3/6/7/8: Lora modem designers guide,” Jul. 2013,
328. revision 1.
 18

[96] W. Du, Z. Li, J. C. Liando, and M. Li, “From rateless to distanceless:

Enabling sparse sensor network deployment in large areas,” IEEE/ACM [119] M. Hata, “Empirical formula for propagation loss in land mobile radio
Transactions on Networking, vol. 24, no. 4, pp. 2498–2511, 2016. services,” IEEE transactions on Vehicular Technology, vol. 29, no. 3,
[97] W. Du, Z. Li, J. Liando, and M. Li, “From rateless to distanceless: pp. 317–325, 1980.
enabling sparse sensor network deployment in large areas,” in Pro- [120] Y. Okumura, “Field strength and its variability in vhf and uhf land-
ceedings of the 12th ACM Conference on Embedded Network Sensor mobile radio service,” Rev. Electr. Commun. Lab., vol. 16, pp. 825–873,
Systems, 2014, pp. 134–147. 1968.
[98] A. Saifullah, M. Rahman, D. Ismail, C. Lu, R. Chandra, and J. Liu, [121] P. E. Mogensen, P. Eggers, C. Jensen, and J. B. Andersen, “Urban
“Snow: Sensor network over white spaces,” in ACM IPSN, 2016. area radio propagation measurements at 955 and 1845 mhz for small
[99] M. Rahman and A. Saifullah, “Integrating low-power wide-area net- and micro cells,” in IEEE Global Telecommunications Conference
works in white spaces,” in Proceedings of the Third International GLOBECOM’91: Countdown to the New Millennium. Conference
Conference on Internet-of-Things Design and Implementation , 2018, Record. IEEE, 1991, pp. 1297–1302.
pp. 255–260. [122] B. Scholkopf and A. J. Smola, Learning with kernels: support vector
[100] X. Xia, Y. Zheng, and T. Gu, “Ftrack: Parallel decoding for lora machines, regularization, optimization, and beyond. MIT press, 2001.
transmissions,” in Proceedings of the ACM Conference on Embedded [123] M. Bor and U. Roedig, “LoRa transmission parameter selection,” in
Networked Sensor Systems, 2019. Proceedings of the International Conference on Distributed Computing
[101] R. Huang, H. Li, B. Hamzeh, Y. Choi, S. Mohanty, and C. Hsu, in Sensor Systems, 2017, pp. 5–7.
“Proposal for evaluation methodology for 802.16 p.” [124] B. Reynders, W. Meert, and S. Pollin, “Range and coexistence analysis
[102] N. Kouvelas, V. Rao, and R. Prasad, “Employing p-csma on a lora of long range unlicensed communication,” in Proceedings of the
network simulator,” arXiv preprint arXiv:1805.12263, 2018. International Conference on Telecommunications, 2016, pp. 1–6.
[103] M. Levorato, U. Mitra, and M. Zorzi, “Cognitive interference manage- [125] C. J. F. Cremers, Scyther: Semantics and verification of security pro-
ment in retransmission-based wireless networks,” IEEE Transactions tocols. Eindhoven University of Technology Eindhoven, Netherlands,
on Information Theory, vol. 58, no. 5, pp. 3023–3046, 2012. 2006.
[104] W.-H. Kuo, R. Kaliski, and H.-Y. Wei, “A qoe-based link adaptation [126] Y. B. Saied and A. Olivereau, “HIP tiny exchange (tex): A distributed
scheme for h. 264/svc video multicast over ieee 802.11,” IEEE Trans- key exchange scheme for HIP-based Internet of Things,” in Proceed-
actions on Circuits and Systems for Video Technology, vol. 25, no. 5, ings of the Third International Conference on Communications and
pp. 812–826, 2015. Networking, 2012, pp. 1–8.
[105] D. Chang, K. Cho, N. Choi, Y. Choi et al., “A probabilistic and op- [127] Y. B. Saied, A. Olivereau, D. Zeghlache, and M. Laurent, “Lightweight
portunistic flooding algorithm in wireless sensor networks,” Computer collaborative key establishment scheme for the aInternet of Things,”
Communications, vol. 35, no. 4, pp. 500–506, 2012. Computer Networks, vol. 64, pp. 273–295, 2014.
[106] L. Xiong, L. Libman, and G. Mao, “Optimal strategies for cooperative [128] P. Porambage, A. Braeken, P. Kumar, A. Gurtov, and M. Ylianttila,
mac-layer retransmission in wireless networks,” in IEEE Wireless “Proxy-based end-to-end key establishment protocol for the Internet-
Communications and Networking Conference, 2008. of-Things,” in Proceedings of the International Conference on Com-
[107] G. Naddafzadeh-Shirazi, P.-Y. Kong, and C.-K. Tham, “Distributed munication Workshop, 2015, pp. 2677–2682.
reinforcement learning frameworks for cooperative retransmission [129] S.-H. Seo, J. Won, S. Sultana, and E. Bertino, “Effective key man-
in wireless networks,” IEEE Transactions on Vehicular Technology, agement in dynamic wireless sensor networks,” IEEE Transactions on
vol. 59, no. 8, pp. 4157–4162, 2010. Information Forensics and Security, vol. 10, no. 2, pp. 371–383, 2015.
[108] C.-Y. Wan, A. T. Campbell, and L. Krishnamurthy, “Psfq: a reliable [130] S. Agrawal, R. Roman, M. L. Das, A. Mathuria, and J. Lopez, “A novel
transport protocol for wireless sensor networks,” in Proceedings of key update protocol in mobile sensor networks,” in Proceedings of the
the ACM international workshop on Wireless sensor networks and International Conference on Information Systems Security, 2012, pp.
applications, 2002. 194–207.
[109] H. Wen, C. Lin, F. Ren, Y. Yue, and X. Huang, “Retransmission or
[131] K. Christidis and M. Devetsikiotis, “Blockchains and smart contracts
redundancy: Transmission reliability in wireless sensor networks,” in
for the Internet of Things,” IEEE Access, vol. 4, pp. 2292–2303, 2016.
IEEE International Conference on Mobile Adhoc and Sensor Systems,
2007. [132] K. Biswas and V. Muthukkumarasamy, “Securing smart cities using
[110] C. Pielli, C. Stefanovic, P. Popovski, and M. Zorzi, “Joint retransmis- blockchain technology,” in Proceedings of the International Conference
sion, compression and channel coding for data fidelity under energy on High Performance Computing and Communications; International
constraints,” arXiv preprint arXiv:1706.09183, 2017. Conference on Smart City; International Conference on Data Science
[111] Semtech, “LoRaMote user guide,” Jul. 2014. and Systems, 2016, pp. 1392–1393.
[112] O. Iova, A. Murphy, G. P. Picco, L. Ghiro, D. Molteni, F. Ossi, and [133] A. Dorri, S. S. Kanhere, R. Jurdak, and P. Gauravaram, “Blockchain
F. Cagnacci, “LoRa from the city to the mountains: Exploration of hard- for IoT security and privacy: The case study of a smart home,” in
ware and environmental factors,” in Proceedings of the International Proceedings of the International Conference on Pervasive Computing
Conference on Embedded Wireless Systems and Networks, 2017. and Communications Workshops, 2017, pp. 618–623.
[113] C. Hakkenberg, “Experimental evaluation of LoRa(WAN) in indoor and [134] S. Huh, S. Cho, and S. Kim, “Managing IoT devices using blockchain
outdoor environments,” Master’s thesis, University of Twente, 2016. platform,” in Proceedings of the International Conference on Advanced
[114] P. Neumann, J. Montavont, and T. Noël, “Indoor deployment of low- Communication Technology, 2017, pp. 464–467.
power wide area networks (LPWAN): A LoRaWAN case study,” in [135] M. Samaniego and R. Deters, “Using blockchain to push software-
Proceesdings of the International Conference on Wireless and Mobile defined IoT components onto edge hosts,” in Proceedings of the Inter-
Computing, Networking and Communications, 2016, pp. 1–8. national Conference on Big Data and Advanced Wireless Technologies,
[115] A. Augustin, J. Yi, T. Clausen, and W. M. Townsley, “A study of LoRa: 2016, p. 58.
Long range & low power networks for the Internet of Things,” Sensors, [136] R. D. Mayra Samaniego, “Blockchain as a service for IoT,” in
vol. 16, no. 9, p. 1466, 2016. Proceedings of the International Conference on Internet of Things and
[116] N. Semiconductors, “Frdm-17510evb evaluation board user guide,” Green Computing and Communications and Cyber, Physical and Social
Apr. 2017, revision 1. Computing and Smart Data, 2016, pp. 433–436.
[117] Cisco, “Cisco 910 industrial router data sheet,” Aug. 2017. [137] J. Lin, Z. Shen, and C. Miao, “Using blockchain technology to build
[118] N. Sornin, M. Luis, T. Eirich, T. Kramp, and O. Hersent, “LoRaWAN trust in sharing LoRaWAN IoT,” in Proceedings of the International
specification,” LoRa alliance, 2015. Conference on Crowd Science and Engineering, 2017, pp. 38–43.