Cluster Computing

https://doi.org/10.1007/s10586-017-1550-8

A clustering WSN routing protocol based on node energy and

multipath
Zhong-xun Wang1 · Min Zhang1 · Xinglong Gao1 · Wenqi Wang1 · Xiaolin Li1

Received: 24 October 2017 / Revised: 11 December 2017 / Accepted: 14 December 2017

© Springer Science+Business Media, LLC, part of Springer Nature 2017

Abstract
Wireless Sensor Network (WSN) is a self-organizing adaptive network composed of infinite sensor nodes which can collect the
data information, process the data information and transmit mutual data information. However, its node power is very limited.
The main responsibility of its routing protocol is to find the scientific and correct forwarding path for the data transmission using
the least amount of energy. The routing protocol of WSN focuses on energy-first, data-centric and application-dependent.
In this paper, we mainly study cluster routing protocol in the wireless sensor network, and analyze the advantages and
disadvantages of LEACH protocol, pointed out the problem. We aiming at the problems existing in the original LEACH
protocol, the cluster head election, the special node processing and inter cluster routing problem were improved respectively,
and then an improved protocol called LEACH-Impt was proposed. With MATLAB simulation, we compare its performance
with the LEACH in the number of survival points and the transmission efficiency of the data. In the last, we compare the
existing time and work efficiency between the old and new LEACH protocol in the hardware system.

Keywords Wireless Sensor Network · Clustering · LEACH · Simulation

1 Introduction environment, and obtain the data information of the monitor-

ing object without the continuous supply of energy. In view of
Wireless Sensor Networks (WSN, or WSNs) [1] is a very these characteristics, wireless sensor routing protocols focus
hot research direction in recent years, because this technol- on energy consumption priority, data centric and application
ogy has a very excellent performance in harsh environments, related. Routing protocol is the lifeline of the whole network;
unattended, limited resources and other special environment. the existence of routing makes the WSN power consumption
The topology of the WSN is not static, and the energy of each is relatively low with fast addressing conditions scalability
node in the network is also limited. Therefore, it is the most and robust. In order to maximize the role of wireless sen-
critical of the WSN to be able to connect all the sensor nodes sor networks in a variety of environments, routing protocol
and reduce the energy consumption. design must be developed in a variety of directions.
The main responsibility of the routing protocol is to find In the structure, WSN routing protocol [2] can be divided
the forwarding path for data transmission, which involves the into plane protocol and hierarchical protocol, the plane proto-
path optimization between the target node and the forwarding col need to maintain a large routing table which is not suitable
node, and guides the data to be transmitted on the correct path. for large-scale network; hierarchical protocol, also known as
Because of the characteristics of different networks, different distributed protocol or clustering protocol, as shown in Fig. 1.
applications, and different requirements, routing protocols
are also different. Wireless sensor network is different from
other networks, and its main function is to adapt to the harsh
1.1 Several other typical clustering protocols

B Zhong-xun Wang The TEEN protocol [3] in hierarchical routing protocols is

[email protected] the first responsive event driven protocol used in WSN, that is
1 to say, the sensor node sends the collected information to the
Institute of Science and Technology for Opto-Electronics
Information, Yantai University, Yantai 264005, Shandong, sink node only when the observation object in the monitoring
China area has a sudden change. Single hop communication is the

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

cluster head node

sink node

chain head

Base
Station

Fig. 2 The chain structure of the PEGASIS protocol

Fig. 1 WSN clustering diagram reduce the amount of time it takes to build this round. How-
ever, the longer the time, the longer the working time of
the cluster head, and the energy consumption of the clus-
communication way between the member nodes and the head ter head will be exhausted, which will make the energy
nodes of each cluster in the TEEN protocol. consumption die prematurely. And if it is shorter, then the fre-
A cluster member node transfers the collected information quent reconstruction of the cluster will cause a lot of energy
directly to the cluster head instead of forwarding the neigh- consumption and waste. The “round” mechanism of TEEN
boring nodes within the time specified in the cluster head; protocol is mainly completed in two stages: the establish-
the cluster head also transmits the fused information to the ment of cluster and the data processing and transmission of
Sink node in a single hop mode [4–8]. module
Compared with LEACH routing protocol, TEEN routing Next, another clustering protocol called PEGASIS [9] is
protocol uses a clustering algorithm similar to LEACH pro- introduced. This is a protocol that is improved on the basis
tocol, and uses the concept of “round” in routing algorithm. of LEACH. The idea of PEGASIS is to divide the nodes in
The main idea of dynamic clustering routing algorithm: the network into multiple chains instead of dividing them
The core of the dynamic clustering routing algorithm is into clusters as before. In this case, there will be no overlap
the “round” mechanism, which mainly displays in the routing in the cluster as in the LEACH protocol. In the PEGASIS
algorithm, the cluster head nodes are dynamically gener- protocol, there is a prerequisite that each node knows the
ated in the “round”, so they are more flexible in practical location of each other, and that the nearest node is cho-
applications. The main design idea of the algorithm is: first, sen as the next hop. Then the greedy algorithm is followed,
cluster head nodes are generated. In the sensor nodes, several and this method is repeated, so that multiple chains can be
cluster head nodes are randomly generated, and the cluster formed. Only nodes close to sink can be selected as chain
head nodes are different in different “round”. First, when the heads. The nodes on the chain can perform data fusion oper-
cluster head node is generated, the message is sent immedi- ations, and the communication between the chain head and
ately to the node in the network. The sensor node is the first the sink can be made directly or by means of other chain
node of the current round; Next, the generation of member heads.
nodes. In the algorithm, some mechanisms and strategies are The chain structure of the PEGASIS protocol is shown
designed. When the cluster head node is generated, the rest in Fig. 2. In this protocol, the chain head is not periodically
of the network will look for the cluster head node accord- changed. Moreover, when the information is transmitted, the
ing to the designed mechanism and strategy to become its nodes on the chain perform the operation of fusing informa-
member node; Last, the transmission and processing of data tion, and the amount of data transmitted can be reduced, so the
within a cluster. An important function of cluster head nodes power consumption of the PEGASIS protocol is obviously
is to coordinate and distribute data transmission among sen- smaller than the power consumption of the LEACH protocol.
sor nodes in a cluster. Of course, the cluster head node will However, since there are only chains in the PEGASIS, when
initially process the received data and then send it to the sink the chain header fails or dies, it results in the failure of data
node. to be transmitted to the sink node. In addition, if the size of
Each round in the TEEN protocol needs to set a suit- the network is large, then the chain will be long, resulting in
able run-time value. If you set it a little longer, you can a great delay in data transmission [10–13].

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

Fig. 3 LEACH operation Clusters establishment stage Stable stage Next round
process

Intra group framing Send to gateway

Time slot

1.2 The LEACH protocol The CH allocate the time points for each subordinate node
to transfer the data in TDMA mode (transmit slot). In the
The more classic of the hierarchical protocol is the one to stable stage, the CH fuses the data which is received by
be discussed in this paper—the Low Energy Adaptive Clus- each common node, and then forward it to the Base Station
tering Hierarchy, LEACH [14]. It was proposed by MIT’s (BS).
Heinzelman in 2000, in which the cluster head (CH) is held
by every node in turns, so the energy consumption of each
node can be balanced as much as possible. However, the 2 Comprehensive improvement of LEACH
protocol also exist defects such as CH distribution is unrea- protocol
sonable, uneven energy consumption, single-hop selection
and so on. Improved protocol, such as W-LEACH [2], T- LEACH protocol can reduce energy consumption by more
LEACH, LEACH-RA and LEACH-C, etc., have improved than 15% compared to other plane protocol [14]. Later, peo-
the original protocol in different directions [14], but most ple began to study the improvement of LEACH protocol and
of the algorithm still use LEACH distributed method in the put forward many practical improvement directions.
election cluster stage.
The selection of CH node is the core process of the 2.1 The main issues of the LEACH protocol
LEACH protocol, so the concept of “round” is proposed.
Each round consists of two states: the selected state and the First, in practical applications, the partitioning of clusters is
steady state. The beginning of each round is to re-determine extremely uneven. Some clusters are too large, while oth-
the location of the CH, and then CH broadcast their own loca- ers have only one node, which results in a great waste of
tion and information, after that the ordinary node received node energy. According to document [15], the energy con-
information from their nearest CH node then join it. When sumption of the minimal cluster nodes is about 20% higher
all the CH and ordinary nodes have been identified (in the than that of the maximum cluster nodes, and the original
actual application, not all of the head has a subordinate ordi- protocol is not reasonable enough for the nodes join cluster
nary node, not all ordinary nodes can join a CH either), the failures.
network began to enter a stable state, and then enter the elec- Second, the communication between cluster nodes and
tion state once again. The whole process is shown in Fig. 3. BS is single path point to point transmission. If the CH is too
It is such a process of periodic elections until the system’s far from the BS, it needs a large transmission power [16].
energy is exhausted and the network is dead At the same time, in the stable phase, if a CH has sudden
 death due to unexpected circumstances, the original LEACH
T (n) = 0, other wise protocol has no spare CH node, and then the ordinary nodes
T (n) = p
, n∈G (1)
1− p×(r mod 1
p)
in the cluster will still communicate according to the original
assigned time until the end of the round, resulting in a great
In order to determine the CH nodes, each node generates deal of energy waste.
a number between [0,1] randomly. If it is less than T (n) Third, document [17] points out that the ratio of CH node
in formula (1), then the node is selected as CH node, and to common node is better than 1/19, that is to say, when
the rest is ordinary node. In formula (1), p is the ratio the proportion is 5%, the system efficiency is the highest.
of the CH node to the total node in the network, r is In fact, the ratio determine the number of CH nodes based
the number of rounds of the current election, and G is on the number of initial nodes, and there is more topology
the set of ordinary nodes in the remaining 1/p round (Fig. changes caused by the addition of new nodes and the death
3). of existing nodes in actual operation.

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

will directly affect the network lifetime, and the probability

of cluster head nodes can not be too large or too small. It
is assumed that there exists an optimal probability of cluster
head to maximize the lifetime of the network, and the best
proportion of cluster heads is analyzed theoretically.
Suppose there are N network nodes in the network, which
are evenly distributed in the area of M * M, The network
selects R cluster head nodes, and the number of nodes in
each cluster is the same. There should be (N /r ) nodes in
each cluster, and each of them consists of 1 CH nodes and
(N /r − 1) CM nodes. Next, the energy consumption of the
Fig. 4 The first order radio model whole cluster is analyzed to obtain the best proportion of
cluster head, so that the energy consumption of the whole
sensing network is lowest.
2.2 First-order radio model In WSN, the cluster heads are responsible for collecting
data from the members of the cluster and fusing them, and
WSN is mainly composed of these subsystems: energy sup- then forwarding them to the BS Station (Base), also known
ply electronic system, communication subsystem, processing as the SINK. Therefore, the energy of cluster head nodes
subsystem and sensing subsystem, besides some sensor is mainly consumed in the following sections: energy con-
networks will have positioning subsystem. The energy con- sumption of data received by members of (N /r ) clusters,
sumed by the communication subsystem is much higher than energy consumption of data fusion processing, and energy
that of the processing subsystem. The experiment proves that consumption of sending the fused data to the base station.
the energy consumption of data transmission is very large, The energy of the nodes in the cluster is mainly in the pro-
and the energy consumption needed for data processing is cess of sending the data to the cluster head node. The energy
relatively small. Based on this principle, the energy con- consumption of a cluster E cluster is divided into two parts:
sumption model mainly considers the energy consumption the energy consumption of cluster head node E C H and the
of the sensing subsystem and the energy consumption of the energy consumption of cluster members E C M . However, the
communication subsystem. energy consumption of the whole network is the sum of the
Figure 4 is a first-order radio model. In this paper, a energy consumption of each cluster, that is, the following
simple model for sending and receiving signals is assumed formula:
[14], the energy consumption is E elec = 50nJ/bit, and the
   
transmit amplifier has a magnification factor of εamp = N
E net = r E cluster = r E C H + − 1 EC M (4)
100pJ/bit/m 2 , so we can get acceptable NE bo value. At the r
same time, it is assumed that there is some energy loss in the
channel transmission. Thus, when the radio model sends k From the above energy model analysis, it is assumed that
bit information from the D distance node, the radio expands each node communication only sends and receives data of a
to: k bit packet, then energy consumption of each CM commu-
nication:
E T x (k, d) = E T x−elec (k) + E T x−amp (k, d)  
∗
E T x (k, d) = E elec ∗
k + εamp k ∗ d2 (2) E C M = E static + k ∗ estatic + ε ∗ d 2 (5)

Under the same conditions, the receive message radio exten- where D is the distance from CM to CH, and the average
sion is: value of D is (from the document [15]):

E Rx (k) = E Rx−elec (k) 1 M2

d2 = (6)
∗ 2π r
E Rx (k) = E elec k. (3)
And:
 
2.3 CH proportion 1 M2
EC M = E static + k ∗ estatic + ε ∗ (7)
2π r
One of the most important factors that affect the lifetime of a
clustered network is the selection of cluster heads. The prob- CM needs to upload all the CH data to BS and it is assumed
ability of the cluster head node is too large or too small, which that CM can ideally fuse CH data, compressing all received

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

data into 1 packets (k bit data). Then the energy consumption recorded, the multi hop path is selected, and one of the best
of CH can be divided into: (N /r −1) times k bit data received performance is used as the current path, and if the path fails,
energy, the energy consumed by a k bit data transmission, and the data is transmitted from the alternative path.
the energy consumption of data compression fusion process- Many researchers use redundant or network coding to
ing (This energy is consumed by the processor program and implement multi-path routing protocols to ensure reliability
can be incorporated into the energy consumption inherent in and real-time. Traditional WSN multipath routing protocols,
the non-RF circuit section E static ), thus: such as AOMDV, can build multiple paths from terminal
  nodes to sink nodes in the routing discovery phase, And the
N least hop path is used to transmit the data until it is found
EC H = − 1 ∗ Er eceive + E send (8)
r unavailable. AOMDV protocol can reduce the number of re
Er eceive = E static + k ∗ estatic (9) discovery of routing, but lead to other serious consequences,
 
E send = E static + k ∗ estatic + ε ∗ D 2 (10) thus, when using multipath routing, we need to consider not
only the number of hops in the path, the energy of nodes, but
also the distance between nodes [18–22].
where D is the distance of CH to BS, from the above three
formula:
3.1 Related multipath routing algorithm
N
EC H = (E static + kestatic ) + kε D 2 (11) In the research work of predecessors, there is an idea to
r
reserve multiple paths in the improvement of multi-path rout-
And: ing mechanism. When the current path fails, the backup path
   is used and new paths are generated on demand to replace
N N N
E net =r E static + kestatic + kε D +
2
−1 the failure paths, thus maintaining the number of backup
r r r paths unchanged. In this paper, we propose a energy based
  
1 M2 multipath routing algorithm (LEACH-Impt), multiple dis-
× E static + k ∗ estatic + ε ∗ (12)
2π r joint paths are generated from the routing topology of inter
cluster nodes. In data transmission, the current optimal path is
Because of (N /r )  1, thus: dynamically selected considering the path energy consump-
   tion, the route hops and the path residual energy [23–26].
N M2 As shown in Fig. 5, there areN nodes and m completely
E net = 2N (E static + kestatic ) + kε r D 2 +
2π r disjoint routing from source node S to sink node (BS) [17].
(13) For the m routing (P1 , P2 , · · · Pm ), location and energy of
each node, the high priority path must have few hops, small
The derivation of R is: transmission distance and large path residual energy.
Nodes energy consumption parameter on the routing path
N M2 is:
D2 − =0 (14)
2π r 2
⎧
N M ⎪
⎪ E 1 = E 11 + E 12 + ... + E 1k1
r= . (15) ⎪
⎨ E 2 = E 21 + E 22 + ... + E 2k2
2π D
.. (16)
⎪
⎪ .
⎪
⎩
E m = E m1 + E m2 + ... + E mkm
3 A clustering WSN routing protocol based
on node energy and multipath Total transmit and receive energy consumption parameters
E i of article i path is:
Nowadays multipath routing can be divided into opportunis-
tic routing, multi-path parallel routing and single path routing k j
extension. Among them, opportunistic routing is a proba- Ei = E i j , (i = 1, 2, · · · , m) (17)
j=1
bilistic model based routing, in which the intermediate nodes
listen to the packets of the source nodes in various probability The Node Ni remaining energy parameters aw after the com-
forms, thus reducing the number of data forwarding; multi- pletion of the data retransmission is:
path parallel transmission is determined by link resource and
hop number, and the data to be transmitted is dispersed to (a0 − ae ) − Ei j
aw = (i = 1, 2 · · · , m; j = 1, 2, · · · , ki )
pre-selected disjoint paths; more common is the last one— a0
expansion of the single path, when the path discovery is (18)

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

Fig. 5 The completely disjoint N12 N1K1

routing between nodes (S,D) N11
N22 N2K2
P1 N21

P2
N1K1
S Nm1 Nm3 D
Nm2

Defines the energy remaining for path i is: evaluation function f (n i , c j ) of CH is defined as follow
[27]:
(a0 −ae )−E i1 (a0 −ae )−E iki
Ri = a0 ∗ (a0 −aae0)−Ei2 ∗ · · · ∗ a0
k j (a0 −ae )−Ei j (19)
= j=1 a0 , (i = 1, 2, · · · , m) ⎧ E(n i )−min{E n k ∈c j (n k )}
⎪
⎨ f e (n i , c j ) = max{E n k ∈c j (n k )}−min{E n k ∈c j (n k )} ,
If the energy of one node is close to the critical state, the max{E n k ∈c j (n k )} = min{E n k ∈c j (n k )} (22)
⎪
⎩
energy consumption parameter converges quickly, and the f e (n i , c j ) = 0, otherwise
reliability and relative balance of the energy transmission of min{dtoB S n k ∈c j }
the path can be guaranteed. f d (n i , c j ) = (23)
dtoB S (n i )
Combining the total energy consumption and hops of the ⎧ N (n i )−min{Nn k ∈c j (n k )}
path, the routing energy priority parameter is defined as: ⎪
⎨ f c (n i , c j ) = max{Nn ∈c (n k )}−min{Nn ∈c (n k )} ,
k j k j
max{Nn k ∈c j (n k )} = min{Nn k ∈c j (n k )} (24)
⎪
⎩
Ri f c (n i , c j ) = 0, otherwise
Pr (i) = (i = 1, 2, · · · , m) (20) ⎧
E i ∗ Hhop
⎨ f (n i , c j ) = we f e (n i , c j ) + wd f d (n i , c j )
+wc f c (n i , c j ), n i ∈ G, we + wd + wc = 1 (25)
Among them, Hhop represents the hop number of S → D, ⎩
f (n i , c j ) = 0, otherwise
and Ri represents the residual energy of path i, and E i repre-
sents the energy consumption of path i. Substituting formulas
(17) and (6) we can obtain the energy priority parameter of
path i: In the Formula (22)–(25), f e , f d and f c are used to examine
the energy condition of the remaining nodes, the distance
k i(a0 −ae )−Ei j to the BS and the number of adjacent nodes respectively.
Pr (i) =
1 a0
, (i = 1, 2, . . . , m). we , wd and wc are weight index whose specific values can
 (21)
Hhop ∗ k1i Ei j be adjusted according to network size and application. At the
same time, the new protocol specifies that the CH election
3.2 A clustering WSN routing protocol based on program will select an alternate CH node . When the CH is far
node energy and multipath from BS, the multi hop routing is used to select the minimum
path to transmit data to the BS. When the distance is close,
Based on the problems existed in the original protocol, in the CH is directly connected with the communication.
this paper, the improved protocol is proposed called LEACH-
Impt whose process is shown in Fig. 6. After the first round,
the next CH is selected by the BS according to the residual
energy, the distance from the BS, the number of times to be 4 Simulation analysis
elected and the number of neighbor nodes; and set alternate
CH nodes and more scientific and reasonable multi hop inter 4.1 Network parameter settings
cluster routing.
The first round election mainly adopts the method of In the Matlab-R2012 environment, the model is built with its
the original protocol, and define wait time is ts , each Editor tool, and 100 nodes are randomly distributed in the
round duration tr ; E curr ent (n i ) is the residual energy of area of 200 * 200. The BS nodes are located at (100, 200)
node n i ; dtoB S (n i ) is the distance of node n i to BS; N (n i ) position, each round is 20 s, and each time the energy con-
is the total number of nodes in the same cluster on this sumption is 50 nJ/bit. The simulation parameters are shown
node; and suppose node n i is belong to cluster ci . So the in Table 1.

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

Yes No
Is it for the
first round?

BS calculates the
Yes No number of surviving
R(n)<T(n) nodes and the optimal
cluster

Common
CH node
node Satisfy CH
requirement

Yes No
Broadcast CH Wait for the
status broadcast of CH
Yes Add CH Join
nodes common
collection nodes
Receive collection
ts Tr / 2
notice

No BS Find
Add the nearest the
Wait for request to cluster according to division of
the
join a cluster the intensity of the optimal
signal cluster

Crate TDMA Waiting schedule BS broadcasts the roles of

timetable notification each node

Receive the
No
timetable?

Yes Adding a
cluster failed
to become a
Ts cleared special node

Transmission data in
stable state

Next round

Fig. 6 LEACH-Impt operation process

Table 1 The simulation parameters 4.2 Simulation result analysis

Item Parameters
The program starts running. As shown in Fig. 7, the dis-
The numbers of nodes 100 tribution density of the nodes is relatively average, without
Distributed area 200 × 200 m excessive concentration or sparseness. The CH nodes within
BS location (100, 250) the BS 75 m are required to communicate directly with the
Initial node energy 1.0 J BS, and the other CH take multi hop routing.
Per round time 20 s The simulation system uses LEACH-Impt optimization
Transmit and receive circuit consumption 50 nJ/bit algorithm for clustering and communication. The selection
Message length 4000 bit of the first round cluster is the same as that of the original
protocol, but a more reasonable multi hop transmission is
added to the communication among the clusters. The protocol
specifies that the cluster head nodes within the base station

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

250
network diagram, we can find that the CH of the improved
algorithm is more uniform, and the multi-hop among clusters
is more reasonable. As shown in Fig. 9, the entire network
200 has 91 dead nodes which is more than 90%, that means the
network has died and the entire network has no CH nodes
and multi-hop transmission.
150

4.3 Comparative analysis

100 Analysis data show that the original protocol has no death
node in the first to 731 rounds, and the first death node
appeared in the 732nd round. Subsequently, the death node
50 increased dramatically, all nodes were dead to the 1220th
round, and the entire network died; The LEACH-Impt has
no dead nodes in the first to the 783 round, and all nodes
0 are alive, the first death node appeared in the 784nd round.
0 20 40 60 80 100 120 140 160 180 200
Subsequently, the death node increased dramatically, but the
Fig. 7 Initial node distribution number of surviving nodes is always larger than the origi-
nal protocol, and the number of dead nodes is always less
than the original protocol, the first death node appeared in
75 m communicate directly with the base station, and the the 1560nd round, all nodes were dead. Through the curve
other cluster heads take multi hop routing. The best path is contrast, it can be intuitively found that the performance of
selected from the path priority given in the third chapter, as LEACH-Impt in survival is better than the original protocol
shown in Fig. 8a (Figs. 10 and 11)
As shown in Fig. 8b, the death node appears in this round. As shown in Fig. 12 is the network data throughput and
The data show that the node appears in the 784th round, transmission contrast of the two protocols (specify that the
52 rounds later than the original death node of the original amount of communication between each cluster and in the
protocol. One of the dead nodes is the cluster head before, cluster is 1), and it can be intuitively found that the amount
which corresponds to the previous theory (cluster head nodes of data transferred by the improved protocol is much higher
are easy to consume and die too fast). than the original protocol.
The death node appears faster after the first death node From Fig. 13, we can find that the energy consumption
appears. As shown in Fig. 9, the network still has a very sci- of the two protocol network is almost the same in the early
entific self-organizing topology. In multi-hop transmission, term (the 0–500 round) of the network operation phase, and
the direction of hop path is also reasonable. Through the the growth rate is relatively fast; In the medium term (round

Fig. 8 First round and dead node appear network routing topology

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

Fig. 9 Snapshot of network 250 250

operation topology
200 200

150 150

100 100

50 50

0 0
0 50 100 150 200 0 50 100 150 200

250 250

200 200

150 150

100 100

50 50

0 0
0 50 100 150 200 0 50 100 150 200

Although a lot of improvements have been made to the

original protocol, which will increase the overall workload
of the agreement, however, the total energy consumption of
the network is certain, so the total energy consumption varies
little, about 50.44 units.

5 Hardware verification

The one value of an algorithm is to apply it to practice.

In the network node, we select the current popular RF
microcontroller—CC2530 to communicate to PC via USB-
CP2102-UART. Host computer interface is designed by Qt
which is a very smart and Object-oriented C++ editor. The
overall architecture diagram is shown in Fig. 14.
CC2530 requires two crystal oscillators, 32 MHz and
32.768 KHz. The LED circuit is directly driven by the port,
Fig. 10 The entire network is paralyzed
when the port output is low, LED is lit. The LED circuit
on the core board is used to indicate the working state of
500–1000), the energy consumption of the two protocol is the nodes, such as joining the network, sending or receiving
different, and the improved protocol energy consumption is data, etc.. The power circuit is different—the coordinator
relatively less than that of the original protocol, and the rate node of this system uses USB to supply power, and the
of growth of the two protocol is relatively fast; In late (round other nodes use 2 AA batteries to supply power. Reset cir-
1000–1500), the energy consumption of the two protocol cuit can reset the microcontroller, is a typical button reset
is different, the improved protocol energy consumption is circuit
relatively less than that of the original protocol, and the rate In the design of the host computer, Qt is adopted. Qt is
of growth of the two protocol is relatively slow. a well-known cross platform C++ graphical user interface

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

100

90

80 LEACH

LEACH-Improvement
70

60
y(dead)

50

40

30

20

10

0
0 500 1000 1500 2000 2500 3000
x(time)

Fig. 11 Comparison of LEACH and LEACH-Impt death nodes over time

4 60
2.5 x 10

40 LEACH
LEACH-Ipmt
2 LEACH
LEACH-Impt

20
y(consumption)

1.5
y(data)

-20

0.5
-40

0
0 500 1000 1500 2000 2500 3000 -60
x(time) 0 500 1000 1500 2000 2500 3000
x(time)

Fig. 12 Comparison of data transmission between LEACH and

LEACH-Ipmt Fig. 13 Comparison of energy consumption between LEACH and
LEACH-Ipmt networks

application framework, which is very large, not just GUI,

but also can be used to develop non GUI programs, such as these five nodes has the function of wireless lighting, aim
console tools and server. Qt is an object-oriented framework to verify whether the network communication patency (Fig.
that uses special code to generate extensions (called Meta 15).
Object Compiler, MOC), and some macros that are easy to Observation of the host computer shows that LEACH-
extend and allow component programming. Therefore, Qt Impt can form a multi hop path between clusters, and the
is chosen to verify the network host computer, and the Qt death time of the first node is about 4 h and 49 min later
program involved in this chapter is based on the Windows than that of the LEACH protocol. Experiments show that
operating system. temperature information acquisition and light control are
The host computer is mainly in order to obtain the topo- unobstructed (Table 2).
logical structure of network, in order to accurately verify This chapter mainly deals with the hardware implementa-
the network quality, the system adds five temperature acqui- tion of the two protocols respectively. The system uses the
sition nodes, and all the nodes are set to away from the current popular radio frequency chip—CC2530 as network
coordinator in the physical distribution network. And the nodes, take CP2102 (convert of USB to UART serial port)
collected temperature will be sent to the PC; in addition, to communicate with the PC machine, and use powerful Qt

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

Fig. 14 Overall architecture diagram

Fig. 15 LEACH-Impt network diagram

Table 2 comparison of network lifetime to design the host computer interface. The network system
LEACH LEACH-Impt is designed quite scientific.

Network lifetime 8 h, 16 min, 53 s 13 h, 05 min, 17 s

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

 Cluster Computing

6 Conclusion 10. Sharef, B.T., Alsaqour, R.A., Ismail, M.: Vehicular communication
ad hoc routing protocols: a survey. J. Netw. Comput. Appl. 40, 363–
396 (2014)
In this paper, the LEACH protocol is studied and improved, 11. Ad hoc wireless networking. Springer, New York (2013)
and the simulation and hardware verification are carried 12. Conti, M., Giordano, S.: Mobile ad hoc networking: milestones,
out. The experimental results show that the survival time challenges, and new research directions. IEEE Commun. Mag.
and data throughput of the improved protocol network 52(1), 85–96 (2014)
13. Ahlawat, A., Malik, V.: An extended vice-cluster selection
have been greatly improved. There is some shortcom- approach to improve v leach protocol in WSN. In: Third Interna-
ing in this paper, When nodes are separated from the tional Conference on Advanced Computing and Communication
network or join the network, the network forms a new Technologies (ACCT), pp. 236–240. IEEE (2013)
topology network structure, which takes a long time. At 14. Heinzelman, W.R., Chandrakasan, A., Balakrishnan, H.: Energy
efficient communication protocol for wireless microsensor net-
this time, it is easy to miss a lot of important informa-
works. In: Proceedings of the 33rd Annual Hawaii International
tion collection and transmission; in an improved protocol, Conference on System Sciences, Piscataway, USA, pp. 175–187
the node’s location moves can cause the network to para- (2000)
lyze. In terms of mobility of nodes, static location of WSN 15. Heinzelman, W.B., Chandrakasan, A.P., Balakrishnan, H.: An
application-specific protocol architecture for wireless microsensor
sometimes has limited its functions. Fast moving wireless networks. IEEE Trans. Wirel. Commun. 1(04), 660–670 (2002)
sensor networks may be the direction of future develop- 16. Arora, V.K., Sharma, V., Sachdeva, M.: A survey on LEACH
ment. and other’s routing protocols in wireless sensor network. OPTIK
In future work, the issue of the topology changes and 127(16), 6590–6600 (2016)
17. Liu, A., Zheng, Z., Zhang, C., Chen, Z., Shen, X.: Secure and
information delays in the network when the nodes are sud- energy-efficient disjoint multipath routing for WSNs. IEEE Trans.
denly joined and separated; and the self-organization of Veh. Technol. 61(7), 3255–3265 (2012)
the network in the case of high speed movement will be 18. Shi, W., Ling, Q., Wu, G., et al.: Extra: an exact first-order algorithm
focused. for decentralized consensus optimization. SIAM J. Optim. 25(2),
944–966 (2015)
19. Jiang, D., Ying, X., Han, Y., et al.: Collaborative multi-hop routing
in cognitive wireless networks. Wirel. Pers. Commun. 86(2), 901–
References 923 (2016)
20. Meng, T., Wu, F., Yang, Z., et al.: Spatial reusability-aware routing
1. Hassan, F., Roy, A., Saxena, N.: Convergence of WSN and cog- in multi-hop wireless networks. IEEE Trans. Comput. 65(1), 244–
nitive cellular network using maximum frequency reuse. IET 255 (2016)
Commun. 11(5), 664–672 (2017) 21. Tyagi, S., Kumar, N.: A systematic review on clustering and rout-
2. Ding, X., Tian, Y., Yu, Y.: A real-time big data gathering algo- ing techniques based upon LEACH protocol for wireless sensor
rithm based on indoor wireless sensor networks for risk analysis of networks. J. Netw. Comput. Appl. 36(2), 623–645 (2013)
industrial operations. IEEE Trans. Ind. Inform. 12(3), 1232–1242 22. Yao, Y., Cao, Q., Vasilakos, A.V.: EDAL: an energy-efficient,
(2016) delay-aware, and lifetime-balancing data collection protocol for
3. Manjeshwar, A., Agrawal D.P.: TEEN: a routing protocol for heterogeneous wireless sensor networks. IEEE/ACM Trans. Netw.
enhanced efficiency in wireless sensor networks. In: IEEE, p. (TON) 23(3), 810–823 (2015)
30189a (2001) 23. Hoang, D.C., Yadav, P., Kumar, R., et al.: Real-time implementa-
4. Zhu, Z., Lu, W., Zhang, L., et al.: Dynamic service provisioning in tion of a harmony search algorithm-based clustering protocol for
elastic optical networks with hybrid single-/multi-path routing. J. energy-efficient wireless sensor networks. IEEE Trans. Ind. Inform.
Lightwave Technol. 31(1), 15–22 (2013) 10(1), 774–783 (2014)
5. Luo, J., Hu, J., Wu, D., et al.: Opportunistic routing algorithm for 24. Soares, V.N.G.J., Rodrigues, J.J.P.C., Farahmand, F.: GeoSpray: a
relay node selection in wireless sensor networks. IEEE Trans. Ind. geographic routing protocol for vehicular delay-tolerant networks.
Inform. 11(1), 112–121 (2015) Inf. Fusion 15, 102–113 (2014)
6. Zhai, C., Lafferty, J.: A study of smoothing methods for language 25. Liu, J., Wan, J., Wang, Q., et al.: A survey on position-based routing
models applied to ad hoc information retrieval. In: ACM SIGIR for vehicular ad hoc networks. Telecommun.Syst. 62(1), 15–30
Forum. ACM, vol. 51(2), pp. 268–276 (2017) (2016)
7. Al-Sultan, S., Al-Doori, M.M., Al-Bayatti, A.H., et al.: A com- 26. Hinds, A., Ngulube, M., Zhu, S., et al.: A review of routing proto-
prehensive survey on vehicular ad hoc network. J. Netw. Comput. cols for mobile ad-hoc networks (manet). Int. J. Inf. Educ. Technol.
Appl. 37, 380–392 (2014) 3(1), 1 (2013)
8. Zhang, X.M., Zhang, Y., Yan, F., et al.: Interference-based topology 27. Chinchu, T., Sangeetha, C.P., Suriyakala, C.D.: Multi-hop LEACH
control algorithm for delay-constrained mobile ad hoc networks. protocol with modified cluster head selection and TDMA schedule
IEEE Trans. Mob. Comput. 14(4), 742–754 (2015) for wireless sensor networks. In: 2015 Global Conference on Com-
9. Lindsey, S., Raghavendra, C.S.: PEGASIS: power-efficient gath- munication Technologies (GCCT), Thuckalay, India, pp. 539–543
ering in sensor information systems. In: Aerospace conference (2015)
proceedings, vol. 3, p. 3. IEEE (2002)

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Cluster Computing

Zhong-xun Wang was born in Wenqi Wang born in 1992, master

Yantai, Shongdong, 1964. He degree candidate of Yantai Uni-
received the Ph.D. degree in Naval versity, her research direction is
Aeronautical Engineering Institue, mobile image processing and
in 2009. His research interest is Lane detection.
source channel coding in wireless
communication.

Min Zhang Male, born in 1991, Xiaolin Li was born in Weihai,

master degree candidate of Yantai Shongdong, on March 23, 1989.
University, his research direction She received the Ph.D. degree in
is Internet of things technology electrical engineering and com-
and wireless sensor network pro- munication engineering from the
tocol. Harbin Engineering University
(HEU), Haebin, Heilongiang, in
2016. His research interests are
in sparse array processing, sparse
array design, tensor signal pro-
cessing, and filter bank design.

Xinglong Gao the laboratory man-

ager in Yantai University, who has
an in-depth study of LDPC codes
and Internet of things in WSN
fields.

123

Content courtesy of Springer Nature, terms of use apply. Rights reserved.

Terms and Conditions
Springer Nature journal content, brought to you courtesy of Springer Nature Customer Service Center GmbH (“Springer Nature”).
Springer Nature supports a reasonable amount of sharing of research papers by authors, subscribers and authorised users (“Users”), for small-
scale personal, non-commercial use provided that all copyright, trade and service marks and other proprietary notices are maintained. By
accessing, sharing, receiving or otherwise using the Springer Nature journal content you agree to these terms of use (“Terms”). For these
purposes, Springer Nature considers academic use (by researchers and students) to be non-commercial.
These Terms are supplementary and will apply in addition to any applicable website terms and conditions, a relevant site licence or a personal
subscription. These Terms will prevail over any conflict or ambiguity with regards to the relevant terms, a site licence or a personal subscription
(to the extent of the conflict or ambiguity only). For Creative Commons-licensed articles, the terms of the Creative Commons license used will
apply.
We collect and use personal data to provide access to the Springer Nature journal content. We may also use these personal data internally within
ResearchGate and Springer Nature and as agreed share it, in an anonymised way, for purposes of tracking, analysis and reporting. We will not
otherwise disclose your personal data outside the ResearchGate or the Springer Nature group of companies unless we have your permission as
detailed in the Privacy Policy.
While Users may use the Springer Nature journal content for small scale, personal non-commercial use, it is important to note that Users may
not:

1. use such content for the purpose of providing other users with access on a regular or large scale basis or as a means to circumvent access
control;
2. use such content where to do so would be considered a criminal or statutory offence in any jurisdiction, or gives rise to civil liability, or is
otherwise unlawful;
3. falsely or misleadingly imply or suggest endorsement, approval , sponsorship, or association unless explicitly agreed to by Springer Nature in
writing;
4. use bots or other automated methods to access the content or redirect messages
5. override any security feature or exclusionary protocol; or
6. share the content in order to create substitute for Springer Nature products or services or a systematic database of Springer Nature journal
content.
In line with the restriction against commercial use, Springer Nature does not permit the creation of a product or service that creates revenue,
royalties, rent or income from our content or its inclusion as part of a paid for service or for other commercial gain. Springer Nature journal
content cannot be used for inter-library loans and librarians may not upload Springer Nature journal content on a large scale into their, or any
other, institutional repository.
These terms of use are reviewed regularly and may be amended at any time. Springer Nature is not obligated to publish any information or
content on this website and may remove it or features or functionality at our sole discretion, at any time with or without notice. Springer Nature
may revoke this licence to you at any time and remove access to any copies of the Springer Nature journal content which have been saved.
To the fullest extent permitted by law, Springer Nature makes no warranties, representations or guarantees to Users, either express or implied
with respect to the Springer nature journal content and all parties disclaim and waive any implied warranties or warranties imposed by law,
including merchantability or fitness for any particular purpose.
Please note that these rights do not automatically extend to content, data or other material published by Springer Nature that may be licensed
from third parties.
If you would like to use or distribute our Springer Nature journal content to a wider audience or on a regular basis or in any other manner not
expressly permitted by these Terms, please contact Springer Nature at

[email protected]