This document proposes adapting blockchain technology to wireless sensor networks to improve energy consumption and security. It discusses how blockchain is typically resource-intensive for sensors with limited batteries, computing power and memory. The paper proposes minimizing this impact by designing a dedicated hardware circuit based on chaos theory. This circuit would compress sensor data, generate blockchain hashes and encrypt data, reducing the processing load. Simulations show this hardware approach decreases energy consumption in the wireless sensor network implementing blockchain by up to 63% compared to using software algorithms alone.
1. Improving energy consumption: Adapting blockchain with
WSN using Hardware Compression Sensorization
Journal: IEEE Access
Manuscript ID Access-2022-24126
Manuscript Type:
Special Section: Advances on High Performance Wireless Networks for
Automation and IIoT
Date Submitted by the
Author:
06-Sep-2022
Complete List of Authors: Azeez, Maytham S.; Universitat Politecnica de Valencia
Azez, Aqeel Salman; Universitat Politecnica de Valencia
Campelo, Jose Carlos; Universitat Politecnica de Valencia
Pina, Alberto Bonastre; Universitat Politecnica de Valencia
Subject Category<br>Please
select at least two subject
categories that best reflect
the scope of your manuscript:
Vehicular and wireless technologies, Sensors
Keywords: <b>Please choose
keywords carefully as they
help us find the most suitable
Editor to review</b>:
Blockchain, Wireless sensor networks, Communication system security
Additional Manuscript
Keywords:
For Review Only
IEEE Access
2. AUTHOR RESPONSES TO IEEE ACCESS
SUBMISSION QUESTIONS
Author chosen
manuscript type:
Research Article
Author explanation
/justification for
choosing this
manuscript type:
provides our manuscript a solution to power consumption problem
in wireless sensor networks when combined with blockchain
technology. we have used hardware compressive sensing based on
chaos theory as solution ,in addition generate the hash by chaos as
hardware (dedicated circuit) to reduce power consumption .Â
Author description of
how this manuscript fits
within the scope of IEEE
Access:
WSN is defenseless against security threats and must face many
security challenges. Therefore, security in WSN and IoT networks is
a critical and not an easy issue. All security mechanisms require
certain resources for implementation, such as code space, data
memory, and power to provide power to the sensor, but these
resources are very limited in small wireless sensors, so we are
working on adapting blockchain technology to these networks.
Author description
detailing the unique
contribution of the
manuscript related to
existing literature:
Our work tries to provide solution to the problem of the complexities
of the blockchain represented in the creation of hash and didn't
reduce the size of transmitted data (which provides one of the most
important challenges for energy consumption).Â
Page 1 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. VOLUME XX, 2017 1
Date of publication xxxx 00, 0000, date of current version xxxx 00, 0000.
Digital Object Identifier 10.1109/ACCESS.2017.Doi Number
Improving energy consumption: Adapting
blockchain with WSN using Hardware
Compression Sensorization
Maytham S. Azeez, Aqeel Salman Azez, José Carlos Campelo, and Alberto Bonastre Pina.
Instituto ITACA. Universitat Politècnica de València. Camino de Vera s/n. 46022 Valencia,
España.
Corresponding author: First A. Maytham S. Azeez (maytham.aljebory@alkafeel.edu.iq).
ABSTRACT Nowadays, Wireless Sensor Networks (WSN) are widely used for collecting, communicating,
and sharing information in various applications. Due to its limited resources in terms of computation,
power, battery lifetime, and memory storage for sensor nodes, it is difficult to add confidentiality and
integrity security features. It is worth noting that blockchain (BC) technology is one of the most promising
technologies, because it provides security, avoids centralization, and a trusted third party. However, to
apply BCs in WSNs is not an easy task because BC is typically resource-hungry for energy, computation,
and memory. In this paper, the additional complication of adding BC in WSNs is compensated by an energy
minimization strategy, which basically depends on minimizing the processing load of generating the
blockchain hash value, encrypting and compressing the data that travel from the cluster-heads to the base
station to reduce the overall traffic, leading to reduced energy per node. A specific (dedicated) circuit is
designed to implement the compression technique, generate the blockchain hash values and data encryption.
The compression algorithm is based on chaotic theory. A comparison of the power consumed by a WSN
using a blockchain implementation with and without the dedicated circuit, illustrates that the hardware
design contributes considerably to reduce the consumption of power. When simulating both approaches, the
energy consumed when replacing functions by hardware decreases up to 63%.
INDEX TERMS WSN, blockchain, energy, compression, FPGA
I. INTRODUCTION
Wireless sensor networks (WSN) are multi-hop self-
organizing network systems that consists of many low-cost
wireless sensor nodes (SNs) connected through wireless
communications. The purpose lies in supportively
perceiving, collecting, and processing the information of
predefined variables in a monitoring area and transmitting
it to a base station (BS). As a significant part of the Internet
of Things (IoT), WSNs play an important role in many
applications, such as medical and health related-
treatment, national defense, environmental monitoring,
and smart homes. In most cases, the SNs are operated by a
battery and cannot be recharged. Moreover, the SNs may
be positioned in hard-to-reach or inaccessible
environments and are anticipated to stay operating for
several months or years.
Therefore, the mechanism of reduction for the
consumption of energy of these nodes and maximizing the
life cycle network has lately drawn attraction by many
researchers [1]. Most of WSN are defenseless against
security threats and must face several security challenges
[2]. Therefore, security in WSNs and IoT is a crucial and not
easy problem [3]. All security mechanisms require certain
resources for implementation, such as the code space, data
memory, and energy to provide power to the sensor, but
these resources are very limited in small wireless sensors
[4].
One type of security technology is a blockchain (BC). The
security of BC systems is vital for their acceptance by
potential users [5]. BCs have attracted the attention of
researchers, as they believe that this technology will bring
about extraordinary changes and opportunities in the world
Page 2 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
4. 2
of industries. BCs are seen as a very robust technology for
resolving trusted communications in a decentralized
manner. This technology was introduced in 2008 and has
been circulated by the encryption postal group. The basic
power of a BC is decentralization, which allows direct
transactions (point-to-point). It is possible to use this
approach in distributed systems, since trust is not required
for nodes to perform transactions, because a third party is
no longer needed in a BC [6]. The definition of BC is a chain
of valid transaction blocks, wherein each one contains the
hash of a previous block in the BC. After verifying the
transaction, it is broadcasted to the network and added to
every copy of the BC. Clearly, the advantages of BC
technology can be listed as follows [7]:
⮚ Security: Neither the node nor anyone else,
except for the transmitter and the receiver, can
have access to the data that are transmitted via
the BC.
⮚ Removal of intermediaries: The peer-to-peer
nature of a BC requires no intermediaries.
⮚ Immutability: Nothing on the BC can change, and
any confirmed transaction cannot be altered.
⮚ Permanence: A public BC acts as a public ledger. If
the BC remains active, data will be accessible.
⮚ Speed: Transactions are quicker than with a
centrally controlled ledger.
FIGURE 1. Blockchain structure.
Generally, a block contains a timestamp, key data,
previous and current block hash, and a set of
transactions, which is a piece of data representing an
operation that some users want to carry out. Figure 1
illustrates the structure of a block [8]. When executing
the transaction, it is hashed into a code and then
broadcasted to each node. Since thousands of
transaction records can be contained in the block of
each node, the BC uses the Merkle tree function to
generate the final hash value, which is also the root of
the Merkle tree. This final hash value will be recorded in
the block header (the hash of the current block), with
the Merkle tree function, computing resources, and thus
the data transmission can be greatly minimized [8].
Adopting a BC in WSNs is not easy, as it is necessary that
critical limitations are addressed. Most researchers have
agreed that WSN nodes/IoT nodes have 3 major
limitations, comprising 1) battery capacity, 2) computing
hardware [9–10], and 3) memory. This will cause more
complexity in achieving security features in such
systems. Hence, they would not be fast enough to carry
data securely in real-time [11,47]. The above points
represent the challenges faced by WSNs, in general, and
during the process of applying BC technology, in
particular. More importantly, BC technology is one of the
technologies that requires a lot of memory, computing
power, and bandwidth, which are already limited in
WSNs.
From an initial point of view, the primary cause
leading to energy consumption is the process of
transmitting and receiving data; hence, the greater the
volume of data, the greater the energy consumption will
be, because the transmission and reception time will be
longer. In addition, with regards to the distance between
the transmitter and the receiver, if the distance is far,
the energy consumed will be huge. With regards to the
computation, constant processing data contributes to
the consumption of more energy. Data processing
includes the following:
● A creation block
● Mining (obtaining a consensus)
● Hash generation
● The creation of encryption keys
● An encryption process and using complicated
algorithms
Therefore, the first solution for saving energy is to
minimize the data size that will be transmitted. This can
Page 3 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
5. 3
be accomplished by the use of data compression.
Advances in the algorithms of data compression have
become indispensable in the applications of
multimedia and communications. As is well known,
there are two major types of compression algorithms:
lossless compression and lossy compression. Lossless
compression is used for applications that require an
exact reconstruction of the original data, while lossy
compression is used when the user can tolerate some
differences between the original and reconstructed
representations of the data. In many applications, this
lack of exact reconstruction is not a problem. For
example, when storing or transmitting speech, the
exact value of each sample of speech is not necessary.
Similarly, when viewing a reconstruction of a video
sequence, the fact that the reconstruction is different
from the original is generally not important as long as
the differences do not result in annoying artifacts [46].
Due to the limitation of memory and processor speed,
many compression algorithms are unsuitable for
embedded real-time environments, because most of
the compression algorithms are inapplicable on the SN
[12,40]. One reason for this is the size of the
algorithms; for example, the size of bzip2 is 219 KB and
the size of LZO is 220 KB [13]. Therefore, to overcome
these problems, the technologies of energy-saving
compression have become a requirement [14, 15]
being imperative to design an algorithm that is capable
of low-complexity and small-scale data compression for
sensor networks [13]. Compressed sensing (CS) is a
recently introduced simultaneous data sensing and
compression mechanism that performs by sampling
randomly at a sub-Nyquist rate. In this situation, the
sample time signals can be totally recovered when they
are adequately sparsely dispersed in the domain of
frequency [16]. On the other hand, in the information
security field, CS has received much attention due to
the fact that CS can be regarded as a cryptosystem to
attain simultaneous sampling, compression, and
encryption when maintaining the secret measurement
matrix [41]. So, CS can provide secrecy if the sensing
matrix is changed for every measurement and this type
of cryptosystem can be compared with One Time Pad
since the sensing matrix is used one time [42]. In this
work, it is aimed to substitute the sequence of random
sampling by a deterministic chaotic sequence, as the
latter mimics or approximates randomness in only few
steps. This effective approach to semi-random
sampling across chaotic sequences first emerged in the
study by [17]. Chaos advantage is a lower-cost
hardware design. The cause for this is that the
integration center is capable of regenerating a
deterministically transmitted chaotic sequence when
the parameters of chaos are known, while with random
sampling, unstandardized sampling time instants
should be sent to the integration center with the time
of arrival data [18]. A chaos system is a nonlinear
system, which is governed by deterministic equations
and possesses an unstable structure, so that its output
acts randomly in only a few steps under particular
initial conditions and controlling parameters [18]. The
term ‘chaos’ in common terms means ‘disorder’.
However, in the theory of chaos, the term is described
accurately, even if there is no universally accepted
mathematical definition of chaos. Due to the many
intrinsic properties of chaos, such as sensitivity to initial
conditions, broadband and orthogonality, and the fact
that chaotic signals can be created by low power, low
cost, and small area electronic circuits, the use of chaos
in communication has gained a great deal of attention
[19,20]. For previous reasons, Chaos has been the
subject of research in a variety of areas relating to
science and engineering. In recent years, chaos has also
been applied in many attention-grabbing uses in the
field of communications [21]. Furthermore, since chaos
is just a deterministic equation, it is possible to
implement chaos on hardware, rather than random
sequences, which is attributed to the fact that chaos is
only a deterministic equation [18]. In addition,
compressive sensing by chaos achieved a cryptosystem
simultaneous with data compression. lso, use the
parameters of the chaotic system are used as the
secret key in the construction of measurement matrix
and, also the masking matrix [43].
Therefore, in this work, Blockchain has been proposed
to design an intrusion detection system, there are
challenges when using a blockchain with WSN, such as
energy. So, a solution is proposed to reduce the power
consumption associated with implementing a BC
platform in WSNs. This solution is based on a custom
Page 4 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
6. 4
circuit design that operates according to the chaos
theory principle for CS to generate a hash value,
compression, and encryption of the data. Wherefore is
no longer the need to add an algorithm to encrypt data
that become resource consuming in terms of time and
energy. Moreover, the compression process (CS) is
performed by hardware parts, which reduces the
workload of the contract processor.
To demonstrate the benefits of this approach, we will
measure and compare the energy consumption and
lifetime with a WSN provided traditional blockchain
functionality without our dedicated circuit and, also
with WSNs provided blockchain with (CS) software.
The rest of the paper is organized as follows. Section 2
state-of-the-art overviews the research that has
provided solutions for energy consumption in WSN
with blockchain. In Section 3, the proposed system and
implementation of the dedicated circuit using a field-
programmable gate array (FPGA) is presented. Section
4 presents the experimentation and results. Finally, the
paper is concluded in Section 5.
II. STATE OF THE ART
In this section, the most relevant research that has
provided solutions for energy consumption in BC
implementations for WSNs will be reviewed and the most
important aspects that researchers have achieved in this
field will be outlined.
This way, in [22], a secure framework was presented that
used a smart contract (Ethereum) to choose a cluster-
header (CH) that can be combined with the proposed model
for mapping a CH from a BC. They obtained low energy
consumption results when they compared it with the
existing approaches in the WSN network. In addition, they
ensured that the data were securely transferred from the SN
to the sink by the CHs and the integrity of the decentralized
database. This model was divided into clusters, each of
which had a number of nodes, one of which was a master
node that received a large amount of data. The remaining
nodes (cluster) in the network performed several tasks at
once, one of which was sending the same piece of
information to the master node and mining it directly on the
BC. The master node, on the other hand, oversees the
authenticity of the obtained data by comparing it to the
data stored on the BC. These researchers provided a new
vision of security by utilizing many of the benefits of the BC
in order to avoid most of possible security problems.
In [23], the use of BC technology was proposed to defend
against malicious attacks in IoT networks, to transfer
wireless energy using the consensus algorithm, which is a
lightweight algorithm that is based on delegated proof-of-
stake (DPoS), to provide substantial energy loss reduction
with minimal operating overhead. To design this system,
they proposed a new secure, distributed wireless power
transmission architecture that included 2 planes, comprising
an energy plane and a BC plane. In this model, regarding
with the BC plane, the energy transmitters acted as miners
and were responsible for managing the BC instead of
establishing the BC in the entire network, so that the
energy-constrained smart devices had less overhead. They
abstracted 3 types of traffic flows from the operations of the
BC, comprising the transaction flow, consensus flow, and
coin flow. The results obtained the by their proposal
included energy loss reduction, which was attributed to the
choice of high-power transmitters to manage the BC, which
resulted in the time taken for the specified power
transmitters to reach consensus being shorter. Furthermore,
the probability of a transaction confirmation failure among
the specified power transmitters was less. This confirmed
the potential energy security gains in terms of energy loss
using the DPoS-based BC.
In [24], the red deer algorithm (RDA)-based clustering
technique was presented with BC-enabled secured data
transmission, which was named RDAC-BC. The RDAC-BC
technique was first initialized, and then the derived fitness
function was used to create clusters. The fitness function in
the RDAC-BC model was solely based on the capacity, power
density, node density, distance to nearby nodes, and
distance to the sink. Following the completion of the CH
election process, a BC-based encrypted communication
process was initiated. The experimental validation of the
RDAC-BC technique was evaluated in a number of ways, and
the findings were compared to the established methods.
This technique has portrayed better energy efficiency,
obtained minimum energy consumption, and exhibited a
higher network lifetime when compared to other
algorithms, such as the genetic algorithm (GA), and Ant-Lion
optimization (ALO) and Grey-Wolf optimization (GWO)
algorithms. The energy consumption was still huge due to
the large volume of energy consumption in the data
transmission and reception processes, and the processing
was also still huge. This represented a load on the processor
node, which in turn also resulted in the nodes consuming
more energy.
In [25], it was suggested to combine BC technology with a
multiple-collaborative BS to increase the energy efficiency of
Page 5 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
7. 5
WSNs. In this model, each BS maintained a BC to provide
consistency of the stored data and network lucidity. It
allowed the user to access to the status of the permitted
nodes. The user could also verify the reliability of the data
when multiple BSs in the network shared the same data
using BC technology. All the messages that were sent
between the nodes, users, and the BSs were considered as
transactions. This proposal provided improved energy
efficiency by using the clustering approach. Because of the
use of the BC technique, extra resources were not needed in
the proposed approach. This method still suffers from the
size of transmitted data which is one of the most important
challenges for energy consumption. The important
difference is that in our research we combined data
compression technology with blockchain technology, and
this is the fundamental difference between our method and
the methods mentioned. So, this approach did not provide a
solution to reduce the power consumption caused by the
size of data transmitted, which is one of the most important
power consumption challenges in WSNs. Also, this approach
suggested that a cluster key be generated for each group
instead of creating keys for each node of the cluster. In
addition, the generation of keys and hash by software need
power consumption.
In [26], to minimize power consumption and enhance
battery life, a load-balancing multi-hop (LBMH) routing and
private communication BC platform was developed and
tested using a constant bit rate (CBR). In this system, it was
attempted to design a private BC in order to secure data. In
this framework, the sink node was responsible for
authenticating the participating sensors nodes and it stored
the ID of the node, along with the anonymized BC node IDs.
It also stored all of the public keys (PKs) of all of the
participating SNs, to ease the CHs while retrieving the PKs.
All of the CH updated the IDs and PKs of the BC to the sink.
During the route discovery from the CH of the source to the
sink node, the sink stored the reverse route information to
the source. Hence, the sink did not need to re-discover the
route to the CH again, as long as the routes remain valid.
In [27], linear network coding (LNC) for WSNs was
introduced with BC-powered IoT devices to increase the
efficiency in terms of the number of live nodes, packet
distribution ratio, and optimized residual energy. The
architecture of the proposed model included the selection
of an appropriate CH using K-means, encoding using the
LNC, and security through the BC network. In this model, the
gateway was responsible for data reception and forwarding
to the BC network. The WSN network and gateway were
connected through the BS. The data of the sensor were
inserted into the BC network using the consensus
mechanism. The application layer retrieved the data from
the BC network and performed the data analytics to initiate
the appropriate actions in the applications of the smart
cities.
In [28], a BC-based protocol was suggested for WSNs to
achieve service immutability, availability, and network
transparency by using a cooperative multiple-BS network to
minimize the likelihood of network failure triggered by any
attack on the BS and reduce energy consumption, because
most of the energy is lost in transmitting data to the BS.
This research presents solutions to reduce energy
consumption in sensor networks with blockchain
technology. These solutions consist of proposing protocols
to balance energy consumption between network nodes
and some research suggested using cluster heads with
special specifications to extend the lifetime of the network.
Our work provided some advantages, as the data encryption
process occurs synchronously with the data compression
process by the (dedicated circuit), and this will provide us
with energy gain, unlike this work, the encryption process is
accompanied by energy consumption and high computing
load. In addition, we will avoid the keys exchange in the
encryption process and, hash generation as software, and
this method consumes energy. Finally, in our approach data
was compressed unlike in this work, so the data
transmission will cost high power consumption.
However, the previous works didn't present an optimal
solution to the problem of the complexities of the
blockchain represented in the creation of hash and didn't
reduce the size of transmitted data (which provides one of
the most important challenges for energy consumption).
Our work tries to provide solutions to these challenges.
III. PROPOSED SYSTEM
In this section, the dedicated circuit to provide a solution
when implementing a BC with a WSN is described. This
circuit will be in charge of performing data compression and
security functions (blockchain related functions) to reduce
the load on the nodes’ processors and reduce energy
consumption. The compression process uses the Signal
compression technique with chaos to select samples, that
consider encryption at the same time, while the
authentication operations include generating the hash.
IV. Operation of the Blockchain-WSN with the dedicated
circuit
This subsection describes the basic operation of the
blockchain-WSN on which we will evaluate our proposal.
The designed circuit will be included in its operation (WSN
nodes) to obtain results. Basically, the WSN architecture
Page 6 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
8. 6
follows the well-stablished proposals in this area. This way,
the WSN architecture follows a structure based on clusters,
with cluster heads (CH) and sensor nodes (SN). Several
routing protocols can be used [32].
The BC paradigm is proposed as a communication
procedure among CHs, as well as between CHs and BS in a
WSN. Nevertheless, the data transmission from SNs to their
associated CHs does not involve BC technology. To reduce
complexity and save energy, as it requires high computation,
which consumes a lot of energy, it is suggested that a
protocol such as Leach protocol should be used to select CHs
dynamically [33,34,35], as shown in Figure 2.
FIGURE 2. Network topology
The CHs create a list of the associated sensor IDs to begin
data collection. The SNs send compressed data to the CH
to reduce its size and encrypt it using the CS. The CH
receive this compressed data. The CH waits until it has
gathered all of its associated node values and then builds a
block of a maximum of 1000 bits to reduce the processing
load in CH. If the number of associated nodes of the CH is
N, the CH waits until it receives the N reading, one from
each node in its domain. If the CH receives N readings, it
assumes that the data are ready to form a block, as all
associated nodes have already sent their data. However, if
one or more of the nodes fails to send its reading for any
reason, the CH will form a block after a specific waiting
time. Any node that has no CH associated with it generates
a block and transmits it to the BS directly. If the data are
sent by another CH, it is passed (because the data has
already been compressed and encrypted by SNs), as shown
in Figure 3.
FIGURE 3. Flowchart of the data flowing into the network
The BC technology addition to the WSN provides security
and de-centralization aspects that will empower the WSN
protection but increases energy consumption and
complexity. The BC technology is supported by Leach
protocol that assumes that all of the parties in the
blockchain network are able to transmit encrypted
information ‘blocks’ to the others. However, the block
information will not be considered as reliable unless the
majority of the network partners decide that it is authentic.
In the authentication process, a considerable group of the
network members have to recognize the current hash of the
block and the previous hash keys, which were included in
the block. Each block contains data and initial parameters
used in the chaos equation to generate the hash as shown in
Figure 5. If the network nodes agree by majority (50%+1)
that means the block is authentic, it will be added to chain.
That happens when CHs exchange acceptance/rejection
Page 7 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9. 7
messages among each other. Each node in the network has
its own ledger to store the authentic blocks. Hence, each
node is able to check if the current block has the necessary
information about the previous circulating blocks in the
network. Inserting a block in the network is like a bid that
the network member has to make, as this bid has to wait for
the decision of other members to accept or reject it.
The CH bids a new block of information, and the consensus
algorithm (POS) is used for authenticity [36]. To prevent bids
from overlapping, the BS creates a schedule of bids sent by
the CHs and works according to the principle of first come,
first serve instead of the mechanism of traditions POS.
The consensus algorithm is applied to the consensus
depending upon the proposed authentication criteria, as
shown in Figure 5.
The CH, which creates a new block, will send it to all CH
node neighbors. Each CH node transmits the new block to all
of the SNs in its cluster. As a result of the second stage, the
current version of the BC will be stored on all of the nodes.
This is because each node is allowed to be a CH in other
rounds.
BC security is based on 2 phases. The first comprises the
encryption phase, which is based on CS. The second
comprises transaction authenticity. In the authentication
phase, as shown in Figures 4 and 5, the dedicated circuit
generates a hash value, and it is also responsible for the
authentication of the new block. Figure 4 shows that when
Sn senses the data, the circuit creates a chaotic sequence,
depending on the parameters, to start the sampling process
to compress and encrypt the data simultaneously. When
receiving the compressed and encrypted data, the block Tx
is generated by CH producing the chaotic hash code from
the dedicated circuit, and then transmitting the control
parameters within the new block, as shown in Figure 4.
FIGURE 4. Security system model: hash creation and block generation
At the receiver, the hash value is re-generated by
entering the message and the control parameters CP that
are explained in more detail in the next section (L, α, and X0)
that are sent within the block to a dedicated circuit to re-
generated hash value as in [38], and the resulting hash value
is compared with the current hash value. Hence, the
parameters are entered into the dedicated circuit, the
sequence chaos is generated again, and the hash is
regenerated to determine the authenticity of the block. In
addition, the hash value of the previous block, which was
included in the current block, enables the other nodes to
determine the source authenticity by comparing the
‘previous’ hash with their own ledgers. If both tests are
passed, then the node will vote to accept the current block.
If a node accepts a certain block after the authenticity
process mentioned above, it will add it the chain stored in its
ledger. Therefore, it will be considered as a ‘previous’ block
in the next round. Once accepted by the block, it will be
added to the BC in the ledger of each node, as shown in
Figure 5.
FIGURE. 5. Security system model: authentication process
V. Circuit design and implementation
A. Signal Compression (CS)
The first function of the designed circuit is to compress the
data based on the chaos theory for sampling. The chaos
system is a non-linear system that is governed by
deterministic equations and has an unstable structure, so
that its outputs behave randomly in only a few steps under
certain initial conditions and control parameters [18]. The
advantage of chaotic equations is that the hardware design
is less expensive because it can regenerate the transmitted
deterministic chaotic sequence if the chaos parameters are
Page 8 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
10. 8
known [18]. So, it can be easily implemented on embedded
devices [17]. The control parameters consist of three values
(the length of the chaotic sequence L, α, and the initial
value, X0). These chaotic equations are created through a
simple chaos system, comprising the logistical map
determined by Eq. (1).
(1)
The sampling rate is reduced based on the chaotic sequence
(of the original signal) by matching the sequence numbers
resulting from the chaotic map and the signal indicator to
determine which samples should be sent as a compressed
signal according to the following algorithm:
Algorithm 1 Signal compression using chaotic logistic map
Input: Input L, α, b, d , sos ,v[], x[0]
Output: Y (compressed data)
1: for i= 1 : L-1
2: x[i] ← α x[i-1] *(1-x[i-1])
3: end for
4: for i = 0 : L − 1
5: x[i] ← b* (x[i]+d)
6: end for
7: i=0
8: while (i<sos)
9: for j = 0 : L − 1
10: i ← i *x[j]
11: if (i<=sos)
12: Y← v[i]
13: end if
14: end for
15: end while
Here, n = 0,1, ........, L-1 (initial X0 value), L represents the
length of the chaotic sequence, α is the control parameter, d
and b are constant set to the maximum required sampling
interval [18], SoS represents the length of the original signal,
v is the input signal, and Y represents the value of the
resulting compressed data.
In order to reconstruct the signal, the sending party must
send that the following parameters (L, α, and initial value,
X0). By inserting these parameters into Eq. (1), the same
chaotic sequence as the original signal is generated.
To reconstructing the compressed signal, linear
interpolation is used to restore the signal to its original
state, as shown in Algorithm 2. Furthermore, interpolation is
the process of estimating the unknown values that fall
between the known values It has been shown in [18].
Interpolation is a method of constructing new data points
within the range of a discrete set of known data points. It is
often required to interpolate, i.e., estimate the value of that
function for an intermediate value of the independent
variable [37].
Algorithm 2 Signal decompression
Input: Input L, α, b, d, sos ,Y, x[0]
Output: V (original data)
1: for i= 1 : L-1
2: x[i] ← α x[i-1] *(1-x[i-1])
3: end for
4: for i = 0 : L − 1
5: x[i] ← b* (x[i]+d)
6: end for
7: i=0
8: while (i<sos)
Page 9 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
11. 9
9: for j = 0 : L − 1
10: i ← i *x[j]
11: if (i<=sos)
12: id← i
13: end if
14: end for
15: end while
16: V← interp {id,Y}
B. Encryption and Hash Value Generation
In order to increase data security, the dedicated circuit will
be doing Compressive sensing (CS). The chaotic compressive
sensing is designed to achieve simultaneous compression -
encryption [39,43]. This encryption has many characteristics,
Security analysis showed this method has a large area to
resist brute force attacks, sensitive to the secret keys, robust
against statistical attacks also, and can use the parameters
of the chaotic system are used as the secret key in the
construction of the measurement matrix and also the
masking matrix [43]. The computational complexities of
these operations are much less than that of the current
mainstream RSA encryption scheme [44].
Compressive sensing can provide secrecy if the sensing
matrix is changed for every measurement. This type of
cryptosystem can be compared with One Time Pad since the
sensing matrix is used one time [42].
In addition, the circuit is also responsible for generating
the hash value that represents the backbone of BC
technology and is also regarded as the main block identifier
in the BC. This way has been proposed because this method
is characterized its low complexity and Theoretical analysis
indicate that the algorithm can satisfy all requirements of
the hash function in an efficient and flexible manner and this
method has higher sensitivity to the original message .it has
a better avalanche effect (any changes in the plaintext affect
the ciphertext) [38,45]. The hash is outlined in the following
Algorithm.
Algorithm 3: generation of hash value [38].
Input: L, α=4, b, d, sos, v [], x[0],M (message)
Output: f(x) hash value
Hash generation
1: for i= 1 : L-1
2: x[i] ← α x[i-1] *(1-x[i-1])
3: end for
4: Divide M into blocks
5: padding
6: for j=0 to n-block
7: f(x)= Mi XOR x[i]
8: end
9: display f(x)
Previous algorithms (1 and 3) where coded in Verilog
languages and synthetized into a FPGA. The ISE Design Suite
14.7 environment was used on a commercially available
hardware FPGA device (Family board: Artix7) [29]. From the
FPGA implementation of the proposed circuit, power
consumption results have been obtained. They are shown in
next section. Algorithms for hash generation and CS have
been also implemented in software to compare the
improvements in power consumption of the hardware ones.
VI. EXPERIMENTATION AND RESULTS
In this section, we analyze our proposal throughput as
well as its efficiency assessment and comparison with other
method without the dedicated circuit (software version).
The simulation and evaluation of the proposed method are
accomplished in MATLAB version 2019b environment using
the results obtained from the circuit implementation in the
FPGA and software energy consumption.
Page 10 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
12. 10
A. Compression Signal results
In our work, Signal compression (CS) is used to reduce
the signal size by sampling from signal and these Samples
are chosen based on a logistic map. CS techniques in this
aspect include those by choose samples in a signal (linear
compression). In figure 6 (1,2,3), a sample signal D(x) was
generated and inserted into the dedicated circuit:
When N=5, 4 and 3.
FIGURE 6. (1,2,3) Compression input signal.
In figure 7 signal G(x) was generated as a random digital
signal, determinate, and discrete and inserted into the
dedicated circuit.
FIGURE 7. Compression input signal
In the CS approach, linear compression approach is used
to transform a signal X (original data) into another form Y
(compressed data). Resulting in compressed signals, the
length of the signal (time period) was random between (500
and 1000) and the number of signal points was 1000.
FIGURE 8. Confidence interval of Compression input signal
Figure 8 shows the confidence interval for the
compression ratio of the proposed circuit: with a confidence
interval equal to 95%, obtained compression is between
52.81% and 49.51%. For a confidence interval equal to 99%,
obtained compression is between 53.49% and 48.83%.
Finally, the mean of compression was 51.1640% with a
standard deviation equal to 0.0784.
Page 11 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
13. 11
All previous results were introduced in the MATLAB
environment, where a blockchain implementation of a WSN
will obtain the benefits of our proposal (next subsection).
B. FPGA results of the dedicated circuit
The power consumption of a dedicated circuit that's
used for compression is shown in Figure 9. It can be seen
that the consumption is about 0.082 W when the signal size
is 752 bits.
FIGURE. 9 Dedicated circuit power consumption.
While the consumption power for hash generation is 0.663
W as shown Figure 10.
FIGURE. 10 Hash power consumption.
To compare with software-implemented functions, hash
SHA256 generation algorithms have been implemented in
an ultra-low power microcontroller. A STM32L476 [30] has
been selected as a representative microcontroller of actual
embedded applications. Data from STM [30] shows a power
consumption equal to 28 µA/Mhz using an external
switched-mode power supply (SMPS) and operating at 24
Mhz. This is the minimum consumption in an operational
stage. With this consumption per Mhz, the best results
(minimum consumption) can be obtained. Additionally,
results area also going to be obtained for a more common
case. This way, we are going to experiment with the
microcontroller running at 80 Mhz (its maximum frequency)
and without external switched-mode power supply. This
could be representative of many actual designs.
In [31] different consumption benchmarks results are
shown. As stated in that application note [31] for an 80 Mhz
frequency, a consumption about 120 µA/Mhz represent a
real situation. As there is a linear relationship between the
software function execution time and the energy, the
execution time of those algorithms has been computed and
then the consumption of these algorithms has been
obtained. This way it will be possible compare a system with
(software-implemented) and without (hardware-
implemented) those algorithms to show the benefits of the
dedicated circuit. Therefore, comparing software and
hardware depends on adding/eliminating the complexity of
the computation time and its associated power
consumption. That is, when functions are not carried out by
hardware, software complexity will be higher. Table 1 show
the consumption of those software-implemented
algorithms:
TABLE 1.
Power consumption for SHA256 software algorithm
24 Mhz with SMPS 80 Mhz without SMPS
Algorithm Execution
time
Consumption Execution
time
Consumption
SHA
algorithm
(3800 bits
of input
data)
120 ms 0,0002611 J 36 ms 0,00114998 J
Page 12 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
14. 12
Additionally, the study all the possibilities, a software
compression technique has been designed also in the
software-only version to analyze its benefits. This way, a
software version of the Chaos compression technique
implemented in hardware has implemented in the same
microcontroller. So, a comparison between software with
and without compression techniques, in addiction to
comparing them with the hardware design will be done.
TABLE 2.
Power consumption for software compression - Chaos
24 Mhz with SMPS 80 Mhz without SMPS
Algorithm Execution
time
Consumption Execution
time
Consumption
CS - 1000
input data
( x 16 bit)
2,445 ms 5,42203E-06
J
734 µs 2,34469E-05
J
VII. Results
This section shows the improvements in energy
consumption of wireless sensors nodes with our circuit that
performs the after mentioned blockchain operations. Our
proposal will be compared with a standard implementation
of a WSN with blockchain operation. To obtain the results,
the Matlab tool will be used. With this tool a WSN will be
configured and simulated, and we obtain data
implementation by this simulation. This WSN will use the
well-known Leach algorithm [34,35] to coordinate and
maintain the cluster structure of the network and determine
the routing strategies. To fulfill the Matlab WSN model, data
obtained from the FPGA circuit implementation about its
power consumption will be used. And to compare with a
pure software implemented blockchain operation, the
energy consumption of a software version of blockchain
operations (generation hash) and the chaos compression
technique will be also introduced to the model. This way we
are going to compare a software version of the blockchain
operations in a WSN with one with the dedicated circuit
designed. Additionally, although not common in WSN
application with blockchain, a software version with
compression techniques (in addition to the blockchain
function) will be also evaluated.
This way, in order to obtain the energy consumption of
the nodes, the next steps will be taken into account:
Each SN reads input data and compresses data, then it will
try to send the data to is CH.
The CH collects data from SNs and generates a hash to
form a block. The node’s energy is re-calculated by
subtracting the transmitter energy consumption from the
initial energy, so that the new node energy is [34]:
𝐸𝑛𝑒𝑤
′
=
𝐸𝑜𝑙𝑑 − 𝐸𝐷𝐴 + 𝐸1 × 𝑁𝑏𝑖𝑡𝑠 + 𝐸𝑓𝑠 × 𝑁𝑏𝑖𝑡𝑠 × 𝑑2
𝑖𝑓 𝑑 <
𝐸𝑓𝑠
𝐸𝑎𝑚𝑝
𝐸𝑜𝑙𝑑 − 𝐸𝐷𝐴 + 𝐸1 × 𝑁𝑏𝑖𝑡𝑠 + 𝐸𝑎𝑚𝑝 × 𝑁𝑏𝑖𝑡𝑠 × 𝑑4
𝑖𝑓 𝑑 ≥
𝐸𝑎𝑚𝑝 1
𝐸𝑎𝑚𝑝 2
(2)
Where EDA is the data sensorization energy, and it is only
consumed on the transmitter node, as it collects data using
its input peripherals. E1 is the energy consumed by the
transmit and receive operations and depends on factors
such as the signal coding, filtering or modulation. Efs (free
space) and Eamp (multi-path) fading channel models
depends upon the transmission distance d. If the distance is
below a threshold, the free space model is used. In other
case, the multi path model is chosen.
In addition, we will add the consumption of the dedicated
circuit to fulfill the data compression and hash computation.
This way, in this “hardware” version it is assumed that the
node uses a separate hardware circuit to implement hash
generation and compression, while basic utilities are
performed by the microcontroller. The consumed energy at
each node is then calculated as follows:
(3)
Page 13 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
15. 13
where E’new is the energy of the node after deducting
the basic utilities consumption as in equation 2 above.
For comparison purposes, the hardware hash
computation and chaos compression has been implemented
in software, and their energy consumption have been
obtained as reflected in Table 1 and 2 for different
microcontroller configurations. This way, for software
implementations:
(4)
A. MATLAB simulation scenarios
The MATLAB simulation scenarios used in the research have
the following assumptions:
• SNs were distributed in a random way in a square
100m×100m region.
• SNs were characterized by being homogeneous
they have been assigned a matchless identifier
number in the network, and their energy is limited.
• The BS is located in the center of the square with a
static position.
• All nodes are able to perform a direct
communication with the BS in a single hop.
• The power of the wireless transmitter could be
adjusted. Table 3 lists the basic parameters.
TABLE 3.
Simulation environment parameters.
Parameters Value
Network area size 100 × 100
Maximum Round 2209
Nodes number 100
Initial energy 0.5 J
Maximum block size 1000 bits
Eamp (Energy of amplifier) 0.0013 pJ/m2
Eelec (energy consumed to transmit or receive) 50 nJ/bit
Efs(amplification coefficient of signal) 10 pJ/bit/m2
Dedicated circuit (compression signal) 1.5530e-12 J/bit
Dedicated circuit (hash generation) 2.32e-09 J/bit
Several simulations have been performed to analyze
the performance improvement of the proposed solution
compared with software approaches. This performance
centers on two main parameters: Network lifetime and
total energy consumption.
B. Network Lifetime
In this research, the definition of network lifetime can
be described as the time from the start of the simulation
to the time when the energy in the battery of the last node
does not allow its operation (the node “dies”). In WSNs,
the lifetime of the network can be classified into 2 periods,
as the stable period and unstable period.
• Stable period typically refers to the lapse of time
between the start of the simulation and the time
when the first node dies.
• Unstable period refers to the time from stable period
to the end of the simulation.
In the Leach protocol, the CHs are not only responsible for
communication with the BS but are also responsible for
integrating the data and creating a block. This method can
balance the energy load of the CH, prevent the early death
of CH, and prolong the lifetime network.
Next figures show the network lifetime for the studied
configuration. The figures illustrate the network lifetime of
an embedded system without the our proposed and with it
(software and hardware as following).
The compression technique (chaos) used is responsible
for part of the improvements in the network lifetime. Thus,
we decided to measure the benefits and drawbacks (extra
consumption) of implementing the chaos compression in
the software. Figure 11 shows that the chaos compression
Page 14 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
16. 14
techniques, even in software version, significantly
improves the network lifetime. This way, the benefits of
compression techniques may outweigh the extra
consumption of the algorithm due to the less data to
transmit. However, it remains to be seen what happened if
the microcontroller works at a higher frequency (and then
with higher consumption).
FIGURE 11. Alive nodes Clock frequency = 24 Mhz in the network with
Chaos compression as software and without software Chaos
compression.
In figure 12 shows the results when the microcontroller
works at 80 Mhz. In this case, with more power
consumption, although the results show the benefits for
the software with compression version, the network
lifetime is far from the results obtained with the dedicated
circuit.
FIGURE 12. Network lifetime. Clock frequency = 80 Mhz Chaos
compression implemented software (SW w/C) version / software version
without chaos
The results of simulation showed that the lifetime of
the network regarding the proposed improved solution
was longer than a lifetime without a dedicated circuit.
When hash computation is performed by software, 100%
of the nodes died at round 1184 while when these
functions and the chaos compression technique are
performed by the hardware, of the nodes died at around
2199. This way the addition of the dedicated circuit, even
operating the microcontroller with the minimum
consumption has important benefits as shown in Fig. 13.
FIGURE 13. Alive nodes in the network with hardware compression and
with software tradition blockchain Clock frequency = 24 Mhz
Figure 14 shows the same results when the
microcontroller works at 80 Mhz. In this case, as the
consumption of the microcontroller is higher, only 765
rounds are obtained for the software only version. This
way, the use of the dedicated circuit for blockchain and
compression functions increases by more than three times
the lifetime of the network.
Page 15 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17. 15
FIGURE 14. Alive nodes in the network with hardware compression and
with software tradition blockchain Clock frequency = 80 Mhz
Figures 15 and 16 shown network lifetime in different
cases (without blockchain technology, traditional
blockchain, and our proposal as software and hardware) in
two states Clock frequency = 24 Mhz and Clock frequency
=80 Mhz.
FIGURE15. Alive nodes in the network with(leach without BC, BC with
hardware compression, BC with software compression, and BC
software tradition) Clock frequency = 24 Mhz
FIGURE 16. Alive nodes in the network with(leach without BC, BC with
hardware compression, BC with software compression, and BC
software tradition) Clock frequency = 80 Mhz
C. Total Energy Consumption
The chaos compression technique was included in the
software version. As it can be seen in figures 17 and 18,
the reduction of data to transmit allows an important
energy saving: 39.7 % when running at 24 Mhz and 52%
when running at 80 Mhz. Obviously, results with the
dedicated circuit are better, but these results highlight the
importance of minimizing the workload to transmit by
means of compression techniques.
FIGURE 17. Energy gain. Clock frequency = 24 Mhz Chaos compression
implemented software (SW w/C) version / software version without
chaos
Page 16 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
18. 16
FIGURE 18. Energy gain. Clock frequency = 80 Mhz Chaos compression
implemented software (SW w/C) version / software version without
chaos
In figure.19 and 20 illustrates the energy gain curve
that refers to the difference between the energy expended
to harvest an energy source and the amount of energy
gained. When the microcontroller works at, respectively,
24 and 80 Mhz. This way, even when the microcontroller
works minimizing it consumption (24 Mhz), the dedicated
circuit allows an energy gain equal to 43 %. If the
microcontroller works at a higher frequency, the benefits
are clear: 63 %, that it is to say, when the software-based
system deaths, the system with the dedicated circuit still
has the 63% of energy.
FIGURE 19. Energy gain. Clock frequency = 24 Mhz
FIGURE 20. Energy gain. Clock frequency = 80 Mhz
D. Energy variance
Regarding the energy variance of the nodes, it was very
high in the traditional software state, while when using the
software compression, the changing energy variance
decreased as shown in Fig. 21 in case Clock frequency = 24
Mhz, as well in case Clock frequency =80Mhz as shown in
fig22.
When the dedicated circuit is replaced to compress
data by the software, it is also shown that the energy
Page 17 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
19. 17
variance of the nodes was very high in the traditional state,
meanwhile,
so, when using the dedicated circuit, the changing energy
variance decreased as shown in Fig. 23 in case Clock
frequency = 24 Mhz and also in case of Clock frequency =
80 Mhz as shown in fig 24.
The hardware solution enabled the WSN to live longer than
the traditional one. Therefore, the energy variance of the
hardware equipped WSN keep going high through the
rounds while the traditional network looses its energy.
FIGURE 21. Energy variance per round. Clock frequency = 24 Mhz
Chaos compression implemented software (SW w/C) version / software
version without chaos
FIGURE 22. Energy variance per round. Clock frequency = 80 Mhz
Chaos compression implemented software (SW w/C) version / software
version without chaos
FIGURE 23. Energy variance per round. Clock frequency = 24 Mhz
FIGURE 24. Energy variance per round. Clock frequency = 80 M
VIII. CONCLUSION
Adopting a BC in a WSN is not easy. It is necessary that
several critical limitations (energy, memory size,
computation load and power consumption) are addressed.
This proposal constitutes a step forward in solving these
main limitations for adopting blockchain technology in
WSN. To overcome these problems, a dedicated circuit was
designed to perform some Blockchain related functions,
such as hash generation etc., and to compress the
information to transmit between nodes. CS is a lossy
compression like most compression methods, but it's
Page 18 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
20. 18
urgently needed because WSNs applications are related to
speed, cost, and battery life of sensors, there is a great need
for this technology even if it is for the loss of part of the
information. This way, this paper evaluates how the
hardware implementation of these complex software
functions contribute significantly to reduce the
consumption of energy of the entire network and extending
its the lifetime.
Hardware signal compression is an effective method to
reduce the data to transmit and, therefore, to minimize the
energy requirements. In the other hand, transferring the
software functions (hash computation) and compression
techniques to a specific circuit reduce both the time and the
energy requirements to compute them. The MATLAB
platform was used to test the new proposed approach. In
order to provide the real operation parameters to the Matlab
simulator, a Verilog implementation of the dedicated circuit
has been done. From this real implementation, energy
consumption has been obtained. The energy consumption
of the software functions performed by the dedicated circuit
has been obtained by implementing them in a representative
embedded microcontroller. So, real energy consumption
parameters have been measured in laboratory conditions.
Main results show a reduction of 63 %.
REFERENCES
[1] Wang J, Tawose OT, Jiang L, Zhao D. A, "New Data Fusion
Algorithm for Wireless Sensor Networks Inspired by Hesitant
Fuzzy Entropy". Sensors (Basel, Switzerland). 2019 Feb;19(4). DOI:
10.3390/s19040784.
[2] T. Kgogo, B. Isong and A. M. Abu-Mahfouz, "Software defined
wireless sensor networks security challenges,", IEEE AFRICON,
Cape Town, 2017, pp. 1508-1513, doi:
10.1109/AFRCON.2017.8095705.
[3] M. Mamdouh, M. A. I. Elrukhsi and A. Khattab, "Securing the
Internet of Things and Wireless Sensor Networks via Machine
Learning: A Survey," 2018 International Conference on Computer
and Applications (ICCA), 2018, pp. 215-218, doi:
10.1109/COMAPP.2018.8460440.
[4] Abdullah Said Alkalbani, T. Mantoro and A. O. M. Tap,
"Comparison between RSA hardware and software
implementation for WSNs security schemes," Proceeding of the
3rd International Conference on Information and Communication
Technology for the Moslem World (ICT4M), Jakarta, 2010, pp.
E84-E89, doi: 10.1109/ICT4M.2010.5971920.
[5] M. Pilkington, “Blockchain technology: principles and
applications,” Research handbook on digital transformations, p.
225, 2016.
[6] Z. Zheng, S. Xie, H. Dai, X. Chen and H. Wang, "An Overview of
Blockchain Technology: Architecture, Consensus, and Future
Trends,", IEEE International Congress on Big Data (BigData
Congress), Honolulu, HI, USA, 2017, pp. 557-564, doi:
10.1109/BigDataCongress.85.
[7] Shu Yun Lim, Pascal Tankam Fotsing, Abdullah Almasri, Omar
Musa, Miss Laiha Mat Kiah,Tan Fong Ang, Reza Ismail “Blockchain
Technology the Identity Management and Authentication Service
Disruptor: A Survey”. International Journal on Advanced Science
Engineering and Information Technology,8(4-2), 2018.
[8] Iuon-Chang Lin and Tzu-Chun Liao, “A Survey of Blockchain
Security Issues and Challenges” International Journal of Network
Security, Vol.19, No.5, PP.653-659 2017.
[9] K. Karimi and S. Krit, "Smart home-smartp hone systems:
Threats, security requirements and open research challenges," in
20 /9 International Conference of Computer Science and
Renewable Energies (ICCSRE). IEEE, 2019, pp.1-5.
[10] Y Yang, L. Wu, G. Yin, L. Li, and H. Zhao, "A survey on security
and privacy issue s in internet-of-things," IEEE Internet of Things
Journal, 4(5), 2017 pp . 1250- 1258.
[11] J. Bugeja , A. Jacobsson , and P. Davidsson , "On privacy and
security challenges in smart connected homes," in 2016 European
Intelligence and Security Informatics Confe rence (EISIC). IEEE,
2016, pp. 172175.
[12] J. Uthayakumar, M. Elhoseny and K. Shankar, "Highly
Reliable and Low-Complexity Image Compression Scheme Using
Neighborhood Correlation Sequence Algorithm in WSN," in IEEE
Transactions on Reliability, 69(4), pp. 1398-1423, 2020, doi:
10.1109/TR.2020.2972567.
[13] N. Kimura and S. Latifi, "A survey on data compression in
wireless sensor networks," International Conference on
Information Technology: Coding and Computing (ITCC'05) -
Volume II, Las Vegas, NV, Vol. 2, 2005, pp. 8-13 doi:
10.1109/ITCC.2005.43.
[14] Tarek Sheltami, Muhammad Musaddiq, Elhadi Shakshuki
“Data compression techniques in Wireless Sensor Networks”
Future Generation Computer Systems 64,2016, pp. 151–162.
Page 19 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
21. 19
[15] D. M. Pham and S. M. Aziz, "An energy efficient image
compression scheme for Wireless Sensor Networks," IEEE Eighth
International Conference on Intelligent Sensors, Sensor Networks
and Information Processing, Melbourne, VIC, 2013, pp. 260-264,
doi: 10.1109/ISSNIP.2013.6529799.
[16] E. J. Candes and M. B. Wakin, "An Introduction To
Compressive Sampling," in IEEE Signal Processing Magazine, vol.
25, no. 2, pp. 21-30, 2008, doi: 10.1109/MSP.2007.914731.
[17] Nguyen, L. T., Phong, D. V., Hussain, Z. M., Huynh, H. T.,
Morgan, V. L., & Gore, J. C. "Compressed sensing using chaos
filters". In: Australasian telecommunication networks and
applications conference (ATNAC), 2008.
[18] Alwan, N.A.S., Hussain, Z.M. "Compressive Sensing with
Chaotic Sequences: An Application to Localization in Wireless
Sensor Networks". Wireless Pers Commun 105, 941–950,2019.
https://doi.org/10.1007/s11277-019-06129-z
[19] Moatsum Alawida, Azman Samsudin, Je Sen Teh,Enhanced
digital chaotic maps based on bit reversal with applications in
random bit generators,Information Sciences,Volume
512,2020,ISSN 0020
0255,https://doi.org/10.1016/j.ins.2019.10.055..
[20] Rhea, B.K., Chase Harrison, R., Aaron Whitney, D., Werner,
F.T., Muscha, A.W., Dean, R.N. Hardware Implementation of
Chaos Control Using a Proportional Feedback Controller. In: In, V.,
Longhini, P., Palacios, A. (eds) Proceedings of the 5th
International Conference on Applications in Nonlinear Dynamics.
Understanding Complex Systems. Springer, Cham.2019.
https://doi.org/10.1007/978-3-030-10892-2_16.
[21] Qtadar Hussain, Amir Anees, Ali Hussain AlKhaldi,
Abdulmohsen Algarni, Muhammad Aslam,Construction of chaotic
quantum magnets and matrix Lorenz systems S-boxes and their
applications,Chinese Journal of Physics, 56(4),2018
https://doi.org/10.1016/j.cjph.2018.04.013.
[22] Benaddi H., Ibrahimi K., Dahri H., Benslimane A. A
Framework to Secure Cluster-Header Decision in Wireless Sensor
Network Using Blockchain. In: Belkasmi M., Ben-Othman J., Li C.,
Essaaidi M. (eds) Advanced Communication Systems and
Information Security. ACOSIS 2019. Communications in Computer
and Information Science, vol 1264. Springer, Cham,2020,
https://doi.org/10.1007/978-3-030-61143-9_17.
[23] L. Jiang, S. Xie, S. Maharjan and Y. Zhang, "Blockchain
Empowered Wireless Power Transfer for Green and Secure
Internet of Things," in IEEE Network,33(6), 2019, pp. 164-171,
doi: 10.1109/MNET.001.1900008.
[24] Gia Nhu Nguyen, Nin Ho Le Viet, A. Francis Saviour Devaraj,
R. Gobi, K. Shankar, "Blockchain enabled energy efficient red deer
algorithm based clustering protocol for pervasive wireless sensor
networks, Sustainable Computing: Informatics and Systems",
Volume 28, 2020.
[25] Swetha Priya T.C., Kanaka Durga A."Clustering-Based
Blockchain Technique for Securing Wireless Sensor Networks. In:
Raju K., Senkerik R., Lanka S., Rajagopal V. (eds) Data Engineering
and Communication Technology". Advances in Intelligent Systems
and Computing, vol 1079. Springer, Singapore,2020.
https://doi.org/10.1007/978-981-15-1097-7_39.
[26] J. Marchang, G. Ibbotson and P. Wheway, "Will Blockchain
Technology Become a Reality in Sensor Networks?," 2019
Wireless Days (WD), Manchester, United Kingdom, 2019, pp. 1-4,
doi: 10.1109/WD.2019.8734268.
[27] N. S. Alghamdi and M. A. Khan, "Energy-efficient and
blockchain-enabled model for internet of things (iot) in smart
cities," Computers, Materials & Continua, 66(3), 2021,pp. 2509–
2524.
[28] Islam, N. Towards a Secure and Energy Efficient Wireless
Sensor Network using Blockchain and a Novel Clustering
Approach. Master Thesis. School. 2018.URL:
https://dalspace.library.dal.ca/bitstream/handle/10222/74098/Is
lam-Nazmul-MCSc-MCS-July-2018.pdf?isAllowed=y&sequence=1.
[29] https://www.xilinx.com/products/design-tools/ise-design-
suite.html , last accessed: October, 12, 2021
[30] ST. STM32L4 MCU series Ultra-low-power with performance.
Available online:
https://www.st.com/content/ccc/resource/sales_and_marketing
/promotional_material/flyer/0a/4f/e9/dc/28/fd/49/9e/flstm32l4.
pdf/files/flstm32l4.pdf/jcr:content/translations/en.flstm32l4.pdf
(last accessed: August, 2nd
2021).
[31] “Design recommendations for STM32L4xxxx with external
SMPS, for ultra-low-power applications with high performance”.
Available online:
https://www.st.com/resource/en/application_note/dm00350156
-design-recommendations-for-stm32l4xxxx-with-external-smps-
for-ultralowpower-applications-with-high-performance-
stmicroelectronics.pdf (last accessed: August 2nd
2021).
[32] Shamsad Parvin and Muhammad Sajjadur Rahim “Routing
Protocols for Wireless Sensor Networks: A Comparative Study”
International Conference on Electronics, Computer and
Communication (ICECC 2008).
Page 20 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
22. 20
[33] Amin Rezaeipanah, Hamed Nazari, Gholamreza Ahmadi “A
Hybrid Approach for Prolonging Lifetime of Wireless
SensorNetworks Using Genetic Algorithm and Online Clustering”,
Journal of Computing Science and Engineering, 13(4), 2019, pp.
163-174.
[34] Chunyao FU, Zhifang JIANG, Wei WEI2and Ang WEI, “An
Energy Balanced Algorithm of LEACH Protocol in WSN” , IJCSI
International Journal of Computer Science Issues,10(1), No 1,
2013.
[35] amin Nazari, "Wireless Sensor Networks Simulation ,
MATLAB Central File Exchange",Retrieved August 29,
2020.(https://www.mathworks.com/matlabcentral/fileexchange/
69778-wireless-sensor-networks-simulation).
[36] H. Cho, "ASIC-Resistance of Multi-Hash Proof-of-Work
Mechanisms for Blockchain Consensus Protocols," in IEEE Access,
vol. 6, pp. 66210-66222, 2018, doi:
10.1109/ACCESS.2018.2878895.
[37] Ahmad, Najmuddin & Deeba, Farah. "Study of Numerical
Accuracy in Different Spline Interpolation Techniques". Global
Journal of Pure and Applied Mathematics. 16. 687-693,2020.
[38] Amin, Mohamed & Faragallah, Osama & Abd El-Latif,
Ahmed. "Chaos-based hash function (CBHF) for
cryptographic applications". Chaos Solitons & Fractals -
CHAOS SOLITON FRACTAL. 42,2009 767-772.
10.1016/j.chaos.2009.02.001.
[39] S. Hsieh, T. Hung, C. Lu, Y. Chen and S. Pei, "A Secure
Compressive Sensing-Based Data Gathering System via Cloud
Assistance," in IEEE Access, vol. 6, pp. 31840-31853, 2018, doi:
10.1109/ACCESS.2018.2844184.
[40] T. Jia, Y. Ju, R. Joseph and J. Gu, "NCPU: An Embedded
Neural CPU Architecture on Resource-Constrained Low Power
Devices for Real-time End-to-End Performance,"53rd Annual
IEEE/ACM International Symposium on Microarchitecture
(MICRO), 2020, pp. 1097-1109, doi:
10.1109/MICRO50266.2020.00091.
[41] Y. Zhang, Y. Xiang, L. Y. Zhang, Y. Rong and S. Guo,
"Secure Wireless Communications Based on Compressive
Sensing: A Survey," in IEEE Communications Surveys &
Tutorials, vol. 21, no. 2, pp. 1093-1111,2019, doi:
10.1109/COMST.2018.2878943.
[42] G. Kuldeep and Q. Zhang, "Revisiting Compressive
Sensing based Encryption Schemes for IoT", IEEE
Wireless Communications and Networking Conference
(WCNC),2020,pp.1-6,doi:
10.1109/WCNC45663.2020.9120785.
[43] Ponuma, R., Amutha, R. Encryption of image data
using compressive sensing and chaotic system. Multimed
Tools Appl 78, 11857–11881,2019,
https://doi.org/10.1007/s11042-018-6745-3.
[44] Z. Yang, W. Yan and Y. Xiang, "On the Security of
Compressed Sensing-Based Signal Cryptosystem," in IEEE
Transactions on Emerging Topics in Computing,3(3), pp.
363-371, 2015, doi: 10.1109/TETC.2014.2372151.
[45] Q. Wu, "A Chaos-Based Hash Function,",International
Conference on Cyber-Enabled Distributed Computing and
Knowledge Discovery, 2015, pp. 1-4, doi:
10.1109/CyberC.2015.13.
[46] Khalid Sayood, "Introduction,Editor(s): Khalid
Sayood,In The Morgan Kaufmann Series in Multimedia
Information and Systems,Introduction to Data
Compression",(5th Edition), Morgan Kaufmann,2018,pp. 1-
10,ISBN 9780128094747,https://doi.org/10.1016/B978-0-
12-809474-7.00001-X.
[47] Gousia Thahniyath, M. Jayaprasad, "Secure and load
balanced routing model for wireless sensor networks",
Journal of King Saud University - Computer and Information
Sciences,2020
Page 21 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
23. 21
MAYTHAM SALMAN AZEEZ received a
bachelor's degree in software engineering
from Al-Mansour University College Iraq in
2007, and MSc.IS degree from Osmania
University, India, 2012, respectively, where he
is teaching in AlKafeel University. He is also a
Ph.D. student of the Universitat Politècnica de
València (UPV), Spain, ITACA Institute. His
research interests include the Internet of Things and wireless sensor
network security.
AQEEL SALMAN AZEZ received a
bachelor's degree in software engineering
from Al-Mansour University College Iraq
in 2007, and MSc.IS degree from
Osmania University, India, 2013,
respectively, where he is teaching in
AlKafeel University. He is also a Ph.D.
student of the Universitat Politècnica de
València (UPV), Spain, ITACA Institute.
His research interests include the Internet
of Things and wireless sensor network security.
José C. Campelo received the M.Sc.
degree in computer engineering and the
Ph.D. degree from the Universitat
Politècnica de València, Spain, in 1993
and 1999, respectively. He is currently
Associate Professor at the Department of
Computer Engineering, Universitat
Politècnica de València, and member of
the Smart Electronic and Networks
Research Group (SENS) of the ITACA
institute. He is currently teaching computer networks and industrial
local area networks. His research interests include wireless sensors
networks, internet of things, embedded systems design, and fault
tolerant systems design.
ALBERTO MIGUEL BONASTRE PINA
received the M.Sc. degree in computer
engineering and the Ph.D. degree from the
Universitat Politècnica de València (UPV),
Spain, in 1991 and 2001, respectively,
where he is currently an Associate
Professor. He is also a member of the
Smart Electronic and Networks Research
Group (SENS), ITACA Institute. He is also
teaching computer networks and wireless sensor networks. His
research interests include the Internet of Things and wireless
sensor networks, being especially devoted to fault tolerance in
chemical analysis systems.
Page 22 of 21
For Review Only
IEEE Access
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60