A Context Aware Cooperative Communication Method in Wireless Sensor Networks

Abstract

In wireless sensor networks, the clustering methods proposed to improve the energy efficiency of a sensor network can lead to severe energy consumption by the cluster heads. Furthermore, in the case where an unbalanced cluster structure is generated, the overall lifetime of the sensor networks is significantly reduced. Moreover, existing cooperative communication methods, for example, CoopLEACH, waste energy because of the unilateral cooperative communication method, which does not consider the distance between the sink and cluster heads. To solve these problems, this paper proposes a novel method that transmits and collects data according to the density of the nodes and the distance between the sink and cluster heads. Through extensive performance studies, we observed that our method is 12% more energy-efficient than CoopLEACH.

1. Introduction

Wireless sensor networks (WSNs) consist of many sensor nodes. Due to the characteristics of the environment where those sensors are deployed, the sensor nodes in WSNs are usually difficult to recharge or replace [1–3]. As an attempt to improve the energy efficiency of WSNs that are deployed in such environments, LEACH [4], which is a clustering-based routing protocol, has been proposed. Generally, LEACH randomly selects cluster heads, and each cluster head generates its own cluster. Each cluster head aggregates data from its member nodes and transmits these data to the sink. Experiencing an unbalanced cluster structure in LEACH is highly probable since it randomly selects the cluster heads. Furthermore, if a cluster is dense, the cluster head may consume a huge amount of energy [5, 6]. The required energy consumption for data transmission can increase as the distance between the sink and cluster heads increases owing to the fading problem [7]. To solve this problem, various cooperative communication methods have been proposed [8, 9]. One conventional cooperative communication based method is CoopLEACH [10], where every cluster has a cooperative cluster head. However, CoopLEACH transmits data without considering the distance between the sink and cluster heads. Thus, the cluster heads must transmit the data through cooperative communication using intercommunication even if the sink and cluster heads are close to each other, which may lead to transmission failure or additional energy loss. To solve this problem, we select the cooperative cluster heads by considering the node density and the distance between the sink and cluster heads. Furthermore, the TDMA schedule and transmission methods vary with the state of each cluster head and its cooperative cluster head. By considering node density and the distance, our proposed method allows more reliable and energy-efficient data transmissions.

2. Preliminaries

2.1. LEACH, LEACH-C

LEACH, which is one of the first clustering methods for wireless sensor networks, changes cluster heads randomly over time in order to balance the energy consumption of nodes. In the clustering method, a cluster head essentially consumes more energy than other nodes. Therefore, LEACH adopts the novel policy described below so that all sensors on the network consume an equal amount of energy. The procedure of LEACH is composed of two phases: the setup phase and the steady phase. In the setup phase, each sensor node chooses a random number between 0 and 1. If the random value is less than the threshold value $pi(t)$ , the sensor node is selected as the cluster head. $pi(t)$ is determined by (1). ( $Ci(t)$ = 0 if node i has already been a cluster head in the most recent ( $r\mathrm{mod}N/k$ ) rounds and 1 otherwise.) Consider $$\begin{array}{cc}& pi\left(t\right)=\left\{\begin{array}{cc}\frac{k}{N-k*\left(r\mathrm{mod}\left(N/k\right)\right)},\hfill & Ci\left(t\right)=\mathrm{1},\hfill \\ \mathrm{0},\hfill & Ci\left(t\right)=\mathrm{0}.\hfill \end{array}\right.\end{array}$$ (1)

Here, r is the current round, and N denotes total number of nodes in (1). The decision of each node to become a cluster head is made by taking into account the suggested percentage of the cluster heads k. Additionally, $N-k*r$ denotes the number of nodes that cannot be a cluster head in the rth round. After the cluster heads are selected, the cluster heads notify all sensor nodes in the network. Once the sensor nodes receive this information, they determine the cluster to which they want to belong based on the RSSI (Received Signal Strength Indicator) of the message from the cluster heads. The sensor nodes inform the corresponding cluster head that they will be a member of the cluster. Subsequently, the cluster heads assign a time at which the sensor nodes can send data to the cluster heads using a TDMA approach. In the steady phase, the sensor nodes transmit the sensed data to their own cluster heads. Each cluster head then aggregates the amassed data and sends the result to the sink. After a certain period of time spent in the steady phase, the network goes into the setup phase again to select new cluster heads. Although the dynamic clustering method of LEACH can increase the lifetime of the system, it also brings extra overhead (e.g., head changes, advertisements) that can offset the gains made in energy consumption. LEACH forms clusters of sensor nodes based on the received signal strength and uses local cluster heads as routers to the sink. However, energy efficiency remains a problem as it pertains to sensor nodes. Many methods have been proposed that have attempted to rectify this problem. LEACH-Centralized [4] (LEACH-C) is one such technique [11, 12]. In LEACH-C, the sink node determines a cluster head by calculating the energy status of the sensor nodes. In LEACH-C, each sensor node sends its energy status and position information to the sink node at the starting point of each round. The sink node selects a cluster head and optimizes network lifetime based on this information. LEACH-C, however, requires extra cost for GPS device and communication.

2.2. CoopLEACH

The MIMO (Multiple Input Multiple Output) method was proposed to get transmit diversity gain in the conventional wireless communication. However, considering the characteristic of wireless sensor network, because it is needed to maintain the node production cost at the lowest level, installing many antennas on each node is almost impossible. Thus, the cooperative communication method was proposed to gain effects similar to those of MIMO. CoopLEACH method enabled increasing energy efficiency by applying cooperative communication to the LEACH method. In the conventional methods, when intercommunication is performed for transmitting data from the cluster head to the sink, energy loss occurs due to fading. In order to resolve this problem, as shown in Figure 1, the CoopLEACH method, when cluster configuration is performed, achieves diversity gain when communicating with the sink by selecting the node that has the minimum communication distance to the cluster head, other than the cluster head, as a cooperative cluster head for cooperative communication. CoopLEACH operates as follows.

Figure 1

System model of CoopLEACH protocol.

Step 1.

Elect cluster heads according to the cluster-head election mechanism of LEACH.

Step 2.

Elect $M-1$ nodes that have minimum communication distances from the cluster heads elected in Step 1 as cooperative cluster heads.

Step 3.

Each cluster head creates TDMA schedule and informs the schedule to cooperative cluster heads and member nodes.

Step 4.

Each member node sends data to the cluster head and the cooperative cluster heads simultaneously using TDMA schedule.

Step 5.

Cluster heads and cooperative cluster heads send their data to sink through cooperative communications.

However, because CoopLEACH assumes that the sink is located far from the sensor field, all the clusters select the cooperative cluster head. Thus, when communicating with the sink, cooperative communication comes to be used utilizing intercommunication without considering distance. This CoopLEACH method has a demerit that it can be hardly applied when the sink is located inside or close to the sensor field. Another problem of CoopLEACH is that it performs cooperative communication without exchanging data of the cluster head and the cooperative cluster head. In CoopLEACH, when transmission distances to the cluster head and to the cooperative cluster head are different from each other, a problem of transmission distance excess to one of them is generated as shown in cluster A of Figure 2, or when data transmission to one of them fails due to problems such as noises or obstacles, the cluster head and the cooperative cluster head come to retain different data as shown in the cluster B. However, because CoopLEACH performs cooperative communication by using only data it retains without considering the cases explained above, it would be hard to get diversity gain for cooperative communication.

Figure 2

Data transmit failed.

3. Proposed Method

The proposed method consists of three main phases: setup, steady, and communication phases. In the setup phase, we build clusters in accordance with LEACH method. Then, in the steady phase, we choose the cooperative cluster heads and generate data transmission schedules. Finally, we send and receive the data based on the basis of the distance between the sink and the cluster heads in the communication phase.

3.1. Setup Phase

At the start of the setup phase, we employ (1), which is used in LEACH, to selected cluster heads, although any clustering method can be employed. After the cluster heads are selected, they advertise to all sensor nodes in the network. Once the sensor nodes receive the advertisement, they decide which cluster they want to belong to on the basis of the RSSI of the advertisement from the cluster heads. The sensor nodes inform the appropriate cluster heads that they will join the cluster.

3.2. Steady-State Phase

Subsequently, the cluster heads assign the time when the sensor nodes can send data to the cluster heads according to the TDMA approach. However, in case a cluster has many member nodes, the cluster head must spend much more energy for data aggregation and transmission. Furthermore, because the member nodes follow the TDMA schedule, the cluster may experience data collisions, and the delays in the data transmit result in wastage of energy. To solve this problem, we select a node with minimum communication distance to a cluster head as the cooperative cluster head for the clusters with member nodes that are more than the node number threshold (NT). Thus, we can minimize the signal loss and energy consumption for the data transmission between a cluster head and a cooperative cluster-head node. In case a cluster has a cooperative cluster head, the member nodes are divided into two groups: the first group transmits the data as a cluster head, and the second group transmits the data as a cooperative cluster head following their own TDMA schedule. Because two nodes that belong to each other's group but have the same TDMA schedule can exist, we employ CDMA to prevent collisions between the two groups.

3.3. Communication Phase

In CoopLEACH, every cluster has a cooperative cluster head. However, when the distance between the cluster head and sink is small, cooperative communication might lead to data loss and waste of energy. To solve these problems, we use cooperative communication only when the distance between the sink and the cluster head is larger than the distance threshold (DT). After all schedules are set, each cluster-head node measures the distance to the sink using the RSSI signal for data transmission. If the distance is longer than the DT, the nearest member node is selected as a cooperative cluster head. If a cooperative cluster head already exists, which was selected in the previous steady phase, the cooperative communication messages are simply sent to the node, instead of choosing the new cooperative cluster-head node. After the clusters have been completed, the cooperative cluster head and the cluster head collect data from the sensor nodes and prepare to transmit the data to the sink. During the communication with the sink, the cluster head and its cooperative cluster head transmit their data to each other to prepare for cooperative communication only if a cooperative cluster head is present in the cluster. Although this process may require additional energy consumption, the energy wastage is negligible because the nearest node to the cluster head is elected as a cooperative cluster head. Moreover, because the radio model can be predefined according to the distance between the cluster head and its cooperative cluster head, the additional energy consumption in this process can be minimized. After the data from the sensor are gathered, the cluster heads and the sink communicate with each other using intracommunication if the distance is less than DT; otherwise, they communicate with each other using intercommunication with the cooperative cluster head. Although the cooperative cluster head is selected when the number of nodes is more than the NT, only the cluster head communicates with the sink using intracommunication when the distance between the cluster head and the sink is less than the DT.

3.4. Determining Optimal Value of DT

Cooperative communication comes to get diversity gain through duplicate transmission of the node. Thus, nodes that perform cooperative communication need to transmit same data for efficient cooperative communication. However, because in the CoopLEACH method each member node transmits and receives data, respectively, without additional procedures and directly sends the data, which could be different from one another, to the sink, it might decrease efficiency of cooperative communication. Thus, in order to prevent this problem, this paper enabled exchanging data of the cluster head and the cooperative cluster head before performing cooperative communication so that they can transmit same data. However, this method consumes additional data exchange energy. Thus, the DT value that enables optimum cooperative communication needs to be calculated considering this additional energy. Consider $$\begin{array}{cc}& E_{\text{ch}}+E_{\text{cch}}+{\text{CE}}_{\text{sink}}<E_{\text{sink}}.\end{array}$$ (2)In (2), $E_{\text{ch}}$ and $E_{\text{cch}}$ are energies consumed when the cluster head and the cooperative cluster head exchange data with each other to perform cooperative communication, $\text{C}{\text{E}}_{\text{sink}}$ is energy consumed when the cluster head and the cooperative cluster head transmit data to the sink by using cooperative communication, and $E_{\text{sink}}$ is energy consumed when the cluster head transmits data to the sink without using cooperative communication. Energy consumption was calculated by applying (2), assuming that the distance between the cluster head and the cooperative cluster head is more than 1 m and less than 10 m, and applying the energy consumption model of (6). As shown in Figure 3, it turned out that when the distance between the cluster head and the sink is more than approximately 86 m, even if additional energy consumption is considered, the amount of energy consumption is smaller when cooperative communication is used.

Figure 3

Energy cost according to the distance between the cluster head and sink.

3.5. Proposed Processing Algorithm

The following is the system model and assumption for the proposed clustering algorithm.

• (1)

A sink node's energy is unlimited and the sink node has knowledge of its own position.

(2)

All nodes are able to measure the approximate distance between nodes based on the Radio Signal Strength Indicator.

(3)

Once sensor nodes are deployed, the positions of the nodes are fixed and energy-constrained.

(4)

Sensor nodes can use power control to vary the amount of transmission power which depends on the distance to the receiver.

Based on the abovementioned equations and assumptions, we can create an order of operations for the proposed method as shown in Algorithm 1.

Algorithm 1: Data transmit processing algorithm.

(1) Cluster construction using cluster set-up method in LEACH

(2) for each  Cluster Head  (CH)

(3)  broadcasts adv. Join message;

(4)  Receive join-request message from non-CH nodes;

(5)  if  join request node number > NT then

(6)   Cooperative cluster head selection  (nearest node);

(7)   Create TDMA schedule (include CCH);

(8)   Broadcasts TDMA schedules;

(9)  else

(10)   Create TDMA schedule;

(11)   Broadcasts TDMA schedules;

(12)  end if

(13)  Measurement of the distance between the CH and sink;

(14)  if  distance > DT then

(15)   if  CCH is not exist then

(16)    Cooperative cluster head selection  (nearest node);

(17)    else

(18)    cooperative communication ready message to CCH;

(19)   end if

(20)   CH and CCH receive data from member node;

(21)   CH and CCH data exchanging;

(22)   CH data transmit to sink with CCH (cooperative communication);

(23)  else if  CCH is not exist then

(24)   CH receive data from member node;

(25)   CH data transmit to sink;

(26)  else

(27)   CH and CCH receive data from member node;

(28)   CH and CCH data exchanging;

(29)   CH data transmit to sink;

(30)  end if

(31) end for

Figure 4 shows how the clusters transmit data after the clustering has been completed. The cluster head of cluster A independently communicates with the sink using intracommunication without the cooperative cluster head because the distance of cluster A is less than the DT and the number of nodes is smaller than the NT. In cluster B, although the cooperative cluster head is chosen because the number of nodes in cluster B is larger than the NT, only the cluster head communicates with the sink using intracommunication because the distance from the sink is less than the DT. Although cluster C has fewer nodes than the NT, a cooperative cluster head is necessary for cluster C because the distance from the sink is larger than the DT. Cluster D has more nodes than the NT, and the distance from the sink is longer than the DT. For both clusters C and D, intercommunication is necessary because intracommunication is not designed for communications between distant nodes. However, because frequent intercommunications lead to significant energy consumption, we use cooperative communication to minimize the energy consumption of the sensor network.

Figure 4

Proposed structure of the cluster method.

4. Energy Model

The physical characteristics of WSNs have led to many studies that focus on maximizing the lifetime of the sensor networks using clustering methods. These studies include LEACH and its many variants [13, 14]. However, in these studies, the cluster heads consume much more energy than the nonhead nodes. Such unbalance in energy consumption among the sensor nodes results in severe limitations of the lifetime of the sensor networks. To address this problem, we adopt the cooperative communication in the present study, which was introduced by [4, 10]. Energy equation is given by $$\begin{array}{cc}& E_{\text{total}}=E_{\text{intra}}+E_{\text{inter}},\end{array}$$ (3)where $E_{\text{intra}}$ is the energy consumed in transmitting and receiving data in a predetermined distance smaller than $d_0$ ; $E_{\text{inter}}$ is the energy used in transmitting and receiving data at a distance longer than $d_0$ . The required $d_0$ here is defined by $$\begin{array}{cc}& d_{\mathrm{0}}=\frac{\mathrm{4}\pi \sqrt{L}{h}_r{h}_t}{\lambda },\end{array}$$ (4)where L is the system loss factor not related to the propagation, λ is the wavelength of the carrier signal, and $h_t$ and $h_r$ are the heights of the transmitting and receiving antennas, respectively, above the ground. We set the parameter to show the same trend as that in [4]. $L=1$ indicates no loss in the system hardware. $h_t=h_r=1.5$  m, and λ is related to the carrier frequency. Thus, we use the 914 MHz carrier frequency λ as follows: $$\begin{array}{cc}& \lambda =\frac{\mathrm{3}*{\mathrm{10}}^{\mathrm{8}}}{\mathrm{914}*{\mathrm{10}}^{\mathrm{6}}}=\mathrm{0.328}\hspace{0.17em}\text{m}.\end{array}$$ (5)Using these parameters, $d_0=86.2$  m.

The $E_{\text{intra}}$ and $E_{\text{inter}}$ energy consumption model in (6) are used; that is, $$\begin{array}{l}{E}_{TX}\left(k,d\right)=E_{\text{elec}}*k+{\epsilon }_{\text{amp}}*k*d^n\\ \phantom{E_{TX}\left(k,d\right)}=\left\{\begin{array}{ll}{E}_{\text{elec}}*k+{\epsilon }_{\text{fs}}*k*d^{\mathrm{2}},\hfill & d<d_{\mathrm{0}}\hfill \\ E_{\text{elec}}*k+{\epsilon }_{mp}*k*d^{\mathrm{4}},\hfill & d>d_{\mathrm{0}}\hfill \end{array}\right.\\ E_{RX}\left(k,d\right)=E_{\text{elec}}*k,\end{array}$$ (6)where $E_{\text{elec}}$ is the energy consumption coefficient of the wireless signal in a bit and k is the number of data bits to be transmitted or received. d is the distance between the transmitter and the receiver. If the distance is less than $d_0$ , the transmit power is attenuated according to the Friis free-space equation. ${\epsilon }_{\text{fs}}$ is the energy required at the transmit amplifier to achieve an acceptable SNR at the receiver as follows: $$\begin{array}{cc}& {\epsilon }_{\text{fs}}=\frac{P_{r_{\text{th}}}{\left(\mathrm{4}\pi \right)}^{\mathrm{2}}}{R_b{G}_t{G}_r{\lambda }^{\mathrm{2}}},\end{array}$$ (7)where $P_{r_{\text{th}}}$ is the threshold value of the received power for error-free reception at the receiver, $G_t$ and $G_r$ are the gains of the transmitting and receiving antennas, respectively, and $R_b$ is the radio bitrate. Assuming a 99 dBm thermal noise floor and a 17 dB² receiver noise figure, the required intracommunication SNR should be at least 30 dB to receive signals with no errors. Thus, the minimum received power $P_{r_{\text{th}}}$ for successful reception is $$\begin{array}{cc}& P_{r_{\text{th}}}>\mathrm{30}+\left(-\mathrm{82}\right)=-\mathrm{52}\hspace{0.17em}\text{dbm}.\end{array}$$ (8)

Therefore, the received power for successful reception of the packet must be at least −52 dBm or 6.3 nW. The other parameters are the same as those in [4] ( $G_t=G_r=1$ , $R_b=1$  Mbps). Thus, introducing these values to (7) yields ${\epsilon }_{\text{fs}}=10$  Pj/bit/m².

If the distance is longer than $d_0$ , the transmit power is attenuated according to the two-ray ground propagation equation. ${\epsilon }_{mp}$ is the energy required at the transmit amplifier to achieve an acceptable SNR at the receiver as follows: $$\begin{array}{cc}& {\epsilon }_{mp}=\frac{P_{r_{\text{th}}}}{R_b{G}_t{G}_r{h}_t^{\mathrm{2}}{h}_r^{\mathrm{2}}},\end{array}$$ (9)where $P_{r_{\text{th}}}$ is the threshold value of the received power. In the intercommunication phase, we consider that the channels are subjected to Rayleigh fading. For the cooperative communication, we use the values used in [4]. Thus we can calculate the energy required at the transmit amplifier for cooperative communication corresponding to the SNR required to achieve a BER of 10⁻⁴. The SNR value capable of maintaining the BER values is found to be 35 dBm when one cluster head exists on the basis of the graph in [10]. Therefore, $P_{r_{\text{th}}}$ is expressed by $$\begin{array}{cc}& P_{r_{\text{th}}}>\mathrm{35}+\left(-\mathrm{82}\right)=-\mathrm{47}\hspace{0.17em}\text{dbm}.\end{array}$$ (10)

The SNR value that can maintain the BER values is found to be 20 dBm when a cooperative cluster head exists. Thus, $P_{r_{\text{th}}}$ is expressed by $$\begin{array}{cc}& P_{r_{\text{th}}}>\mathrm{20}+\left(-\mathrm{82}\right)=-\mathrm{62}\hspace{0.17em}\text{dbm}.\end{array}$$ (11)

Therefore, the minimum received signal strength must exceed −47 dBm or 20 nW when one cluster head exists; it must exceed −62 dBm or 0.63 nW to succeed in receiving the packets when a cooperative cluster head exists. In other words, as a result of the application of each value into (9), ${\epsilon }_{mp}$ is equal to 0.00396 Pj/bit/m⁴ when one cluster head exists; ${\epsilon }_{mp}$ is equal to 0.00013 Pj/bit/m⁴ when a cooperative cluster head exists.

5. Simulation Result

Through simulation tests under various setups, we measured the energy consumption to demonstrate the efficiency of the proposed method. Our simulations were performed using the Network Simulator 2 (NS2) [15]. We used the scenario generator included in NS2 to construct the sensor network topologies. The energy consumption models in our simulations were based on the First Order Radio Model as shown in (6), which is used in the LEACH experiments. The simulation parameters used in [4, 10] are given in Table 1.

Table 1

Simulation parameters.

Environment parameter	Value
Region of sensor network	100 × 100
Sink node position	(50, −150), (50, −50) (50, 0), (50, 50)
Total node number: n	100
Initial energy of each node	2 J
$$E_{\text{elec}}$$	50 nJ/bit
$${\epsilon }_{\text{fs}}$$	10 Pj/bit/m2
$${\epsilon }_{\mathrm{mp}}$$	0.00396 Pj/bit/m4
$${\epsilon }_{\mathrm{mp}}$$ with CCH	0.00013 Pj/bit/m4
Packet size (k)	4000 bits
$$d_{\mathit{0}}$$	86 m
NT	30
DT	86.2 m

We assume a sink and change the position of the sink for each test. The known optimum ratio of the cluster heads to the entire nodes is 5% [4]. Therefore, in case a given sensor network contains 100 nodes, the optimal numbers of clusters and nodes per cluster are 5 and 20, respectively. Thus, if a cluster contains more than 30 nodes, we assume that the node density of the cluster is high (NT = 30). In addition, we set the DT equal to $d_0$ (=86 m).

Figure 5 shows the volume of received data as a function of the average data transmission time for LEACH, CoopLEACH, and the proposed method; it shows that the proposed method receives more data than the others because LEACH and CoopLEACH suffer from delays resulting from the lack of TDMA slot and from data collision. Because LEACH and CoopLEACH methods transfer data regardless of the node density of each cluster, they have more chances of data collisions and fewer TDMA slots. In contrast, the proposed method minimizes the chances because it determines the TDMA schedule considering both the cluster heads and cooperative cluster heads. Thus, the proposed method minimizes the delays resulting from data collisions and lack of TDMA slots.

Figure 5

Average time for data transmission.

Figure 6 shows the number of surviving nodes in each round of LEACH, CoopLEACH, and the proposed method. By minimizing the delays caused by the data collisions and lack of TDMA slots, we achieved up to 2.7 times longer than LEACH and up to 1.12 times longer than CoopLEACH.

Figure 6

Number of surviving nodes over rounds.

6. Conclusion

This paper has proposed a method that employs an efficient cooperative communication method by considering the node density and distance between the sink and cluster heads for WSNs with energy-usage constraint. We observed that the proposed method exhibited up to 1.12 times more energy-efficiency than CoopLEACH. For our future work, we plan to augment our method with existing cluster-head selection to generate more energy-efficient clusters.