Fuzzy Logic Election of Node for Routing in WSNs

Fuzzy Logic Election of Node for Routing in WSNs Shaik Sahil Babu1, Arnab Raha2, Mrinal Kanti Naskar3

Department of Electronics and Telecommunication Engineering Jadavpur University, Kolkata – 700 032, West Bengal, India

Omar Alfandi4 College of Information Technology, Zayed University, UAE.

Institute of Computer Science, SensorLab, Telematics Group, Georg-August-Universität Göttingen, Göttingen, Germany

Dieter Hogrefe5

Institute of Computer Science, SensorLab, Telematics Group, Georg-August-Universität Göttingen, Göttingen, Germany

[email protected], [email protected], [email protected] [email protected], [email protected]

Abstract— Sensor nodes of Wireless Sensor Networks (WSNs) are resource constraints in energy, memory, processing and communication bandwidth. Since they are operated by battery, their life span is limited. Specially, energy conservation is very important issue in the WSN, because it directly affects the life of the node as well as the entire network. Here, we develop a new way of electing a node among many trustworthy nodes for routing processes. This method consumes the energies of network nodes based on Fuzzy logic applied on their residual energy, trust level and distance from the Base Station. The proposed method elects one indispensible node for participating in routing among many worthy nodes. Hence, this method of election of node for routing in WSN sees the conservation of nodes energies go by very smooth and justifying, thereby increasing the life of the WSN.

Keywords- Wireless Sensor Network; Fuzzy Logic Controller; Node; Routing Protocol; Trust Management System; Residual Energy; Base Station.

I. INTRODUCTION

A Wireless Sensor Network (WSN) consists of spatially distributed autonomous wireless sensors (also called nodes or motes) to cooperatively monitor physical or environmental conditions, such as temperature, humidity, light intensity, sound, vibration, pressure, motion etc., as shown in Figure 1. It is a network made of hundreds or thousands of the sensor nodes which are densely deployed in hazardous/unattended environment with the capability of sensing, computing and communicating wirelessly to the Base Station (BS) also called Sink, via neighbor benevolent nodes. The sink can use many ways to communicate with remote network, such as Internet, satellite and mobile communication network. Finally, the Task Manager (User) collects this transmitted data. WSNs suggest solutions in various applications domains such as military fields, healthcare, homeland security, industrial control, intelligent green aircrafts and smart roads as in ([1]-[3]).

Now a days, in WSNs, a new security system is being applied, called trust management system. The traditional

security systems proposed in ([4]-[7]) such as public-key cryptographic, authentication and other mechanisms cannot be applied directly in WSNs because they are computation intensive at sensor nodes with limited memory, battery life, computation and communication capabilities. Also, WSNs are susceptible to a many variety of attacks ([8]-[11]) like node capture, eavesdropping, worm-hole, Sybil attack, sink-hole and denial of service. Hence, it is clear that any new algorithm or methodology that being designed to incorporate in the WSNs, should be light weight in resource point of view.

Also, many researches on routing protocols related in WSN are processed, but it is required to design and develop a light weight routing protocol in energy consumption point of view, and which gives equal priority to all neighbor nodes for electing one node as an indispensible node for routing based on nodes trust level, residual energy and their distance from the base station. All the sensor nodes in sensor field operate on battery which is limited enough. Since the main objectives of the sensor nodes are sensing/collecting events data, processing data, and transmitting processed data through routing to the Sink. The life time of the sensor node and the entire WSN not only depends on the residual energies of the nodes, but also on the energy efficiencies of different communication algorithms used for trust management and routing protocols.

2012 IEEE 11th International Conference on Trust, Security and Privacy in Computing and Communications

978-0-7695-4745-9/12 $26.00 © 2012 IEEE

DOI 10.1109/TrustCom.2012.166

1279

The energy conservation at node level increases the entire network life span.

In this paper, we extend our previous work [12], for evaluating neighbor nodes priority levels and electing one indispensible node among many trustworthy neighbor nodes, based on the node residual energy, distance from the Sink and the trust level. Fuzzy Logic is applied on node’s residual energy, distance from the Sink and the trust level, to find its Priority level for routing. Finally, highest Priority level node will be elected as the appropriate node for routing. We have presented graphically, how the priority level of neighbor node varies with respect to its trust, distance and energy level. Performance results presented for assumed case study. The remainder of this paper is organized as follows: Section II briefly describes the trust model and the importance of Fuzzy Logic as background knowledge. Section III introduces Fuzzy Logic Election of Node for Routing in WSNs in detail. Section IV Simulation Results and Performance Evaluation, and in section V Conclusions.

II. RELATED WORK

A. Nodes Trust Levels The trust definition, evaluation method proposed in [12]. The direct trust of any node, say A on the node B is a geometric mean function of the all trust metrics as shown in Equation 1. Here, (B) is a direct trust of node A on node B. And

where i=1 to k is a set of K different trust metrics. …. (1)

The indirect trust IT on node B with respect to A can be calculated from the direct trusts (DTs) on B with respect to its L neighbor nodes sent to the node A. It is also a geometric function but of direct trusts as shown in Equation 2.

…. (2) The is the indirect trust of node A on node B, calculated for indirectly given information by L neighbor nodes of node B. And where i = 1 to L neighbors, is set of direct trusts of L neighbor nodes on node B and given to node A.

The weights is weightage given to DT and to the IT where + . Weights can be assigned using different approaches. Sometimes DT may be given more weight, and IT may be given less weight i.e. > .

B. Relation between energy dissipation and its travelling distance.

The energy consumption analysis [13], assuming first order radio model, in which the energy expanded to transfer a k-bit packet to a distance d, and to receive that packet, as suggested by H.O. Tan and I. Korpeoglu is:

….(3)

….(4) Here, is the energy dissipation of the radio in order to run the transmitter and receiver circuitry and is equal to 50nJ/bit. The is the transmit amplifier that is equal to 100pJ/bit/m2. In a wireless channel, the electromagnetic wave signal strength falls off as a power law function of the distance between transmitter and receiver. If the distance between the transmitter and receiver is less than a certain cross-over distance, the Friss free space model is used (d2 attenuation), and if the distance is greater than cross-over distance, the two-ray ground propagation model is used (d4 attenuation).

, If d < cross-over distance …(5) , If d > cross-over distance …(6)

C. Fuzzy Logic In any system, when an input data is insufficient to form the output, any one of the following method is best for output formation.

1. Fuzzy Logic 2. Probability 3. Graph Theory

Fuzzy logic is being applied to a wide variety of applications. There are 3 types of fuzzy control systems/models used.

1) Mamdani Fuzzy model 2) Sugeno Fuzzy Model 3) Tsukamoto Fuzzy Model

The most commonly used Fuzzy Inference Technique is the so-called Mamdani method. The Mamdani-style Fuzzy Inference Process is performed in the following steps. Step 1: Defining the inputs and outputs for FLC. We need to define the universe of discourse for all of the inputs and outputs. The range of values that inputs and outputs may take is called the universe of discourse. Step 2: Fuzzify the Inputs Symmetrically distribute the fuzzified values across the universe of discourse. Use an odd number of fuzzy sets for each variable so that some set is assured to be in the middle. The use of 5 to 7 sets is fairly typical. Overlap adjacent sets (by 15% to 25% typically) Step 3: Setup Fuzzy Membership Functions for the Output(s). Step 4. Create a Fuzzy Rule Base. Construct the rule base for the design. These rules usually take the form IF-THEN rules.

1280

Step 5. Defuzzify the Outputs. The last step in Fuzzy logic inference system is defuzzification. Fuzziness helps us to evaluate the rules, but the final output of a fuzzy system has to be a crisp number. The input for the defuzzification process is the aggregate output fuzzy set and the output is a single number. There are several defuzzification methods, but probably the most popular one is centroid technique. This method finds a point representing the centre of gravity of the aggregated fuzzy set.

III. FUZZY LOGIC ELECTION OF NODE FOR ROUTING IN WSNS

In this section we propose a new method of selecting one best node among many qualified nodes for routing data/control packets. The proposed method called Fuzzy Logic Election of Node for Routing in Wireless Sensor Networks, is applied on three parameters of all neighbor nodes of radio range and finds the highest priority level node. The first parameter is node’s residual energy level (EL), second parameter is node’s distance level (DL) from the BS, and the third is the node’s trust level (TL).

A. Assumption 1 Every node uses trust evaluation method [12] and knows the trust levels (TL) of its neighbors of one radio range (m with respect to node N) as shown in Figure 2. It finds the trustworthy neighbor nodes (say A, B, C) based on the trust threshold Tth. If no node is found trustworthy or only few trustworthy nodes are present in its radio range then it increases its radio range from m to n, and finds the new trustworthy nodes again (nodes D, E, F, G, H, I may be added), which is an energy consuming operation. Every node in WSN maintains a database that contains the history related to their neighbors, i.e., trust metrics of each neighbor node, direct trusts, indirect trusts, trusts at different times, residual energies of neighbors, and their distance (based on RSSI) from the BS. Every node knows residual energy levels (EL) of its neighbor nodes and will be stored in the database of the node. These (EL) are received and maintained by the Trust Management System of the node.

B. Assumption 2 The BS sends a test signal (TS) to the all nodes of the network periodically. The period (tp) of this test signal (TS) is not defined and it is application dependent. This test signal (TS) will be received by all nodes. Every node receives the test signal (TS) but with different strengths (say p, q, r) as shown in Figure 2. Based on Received Signal Strength Indicator (RSSI), every node finds its distance or roughly distance level (DL) from the BS. Every node in the WSN, communicate the distance level (DL) information along with the self residual energy (EL) to the neighbor nodes in any one of two different situations, first, when it is asked for indirect trust information and when its distance from the BS is changed.

To increase the life of the node, as well as the entire network, every node in the network must utilize their energy properly. All the time, only energetic nodes must not be selected for routing. Along with the energy level of the node, the node’s distance from the BS and the trust level must also be taken into consideration and must be given equal priority. In this proposed method, we are giving same priority to the first two parameters, and for the third parameter trust level is given lesser priority, because we are applying this fuzzy logic election method on trusted nodes only, i.e. whose trust level is beyond the trust threshold (Tth). Hence, there is no meaning to give equal priority to trust level as given to other two parameters. The relation between Trust Management System (TMS), Routing Protocol, neighbor node’s Database and the Fuzzy Logic Controller (FLC) is shown in Figure 3. The trust management system periodically finds the neighbor nodes of its radio range, and evaluates their trust levels and store them into the database. It also finds selfish nodes and most benevolent nodes. Another important function of trust management system is, it gets the neighbor node’s residual energy levels and their distances from the BS. All these data maintained in the node’s database. Whenever routing protocol wants to send some packet to the BS, it first

B

C

A N

BS

I

H

D

E

G F

pq

r

mn

Fig. 2 WSN nodes with different RSSIs

Database

Trust Management

System

Fuzzy Logic Controller

Routing Protocol

Inference Engine

Defuzzifier

Fuzzy Rule Base

Fuzzifier

Fig.3 Relationship of FLC with database

1281

instructs the trust management system to update the database. Then the TMS in turn, initiates the FLC to find the indispensable trustworthy node from the available information of the database and update the database. Then the Routing Protocol collects the node information of highest priority level that is indispensable node for routing in that moment. If more than one node is showing the same Priority Level, then the node whose Distance from BS is least among them that will be selected for routing because the packet may reach the destination with minimum number of hops (consuming less energy).

IV. SIMULATION RESULTS AND PERFORMANCE EVALUATION

a) Simulation

We have designed the FLC system as shown in the Figure 3. The inputs to the FLC come from the Database module. The inputs are trust (T) levels, residual energy (E) levels and the distances (D) from the BS. The out put of FLC is Priority level (PL) of the nodes and will be stored in the database module only. The Routing Protocol can find the highest priority level node for routing operations. The inputs and outputs for the FLC and their minimum and maximum values are shown in Table 1. Fuzzifying of Inputs and Outputs: We have used triangular membership functions to fuzzify the inputs. For different inputs the fuzzy variable and its crisp input ranges are shown in below Table 2. The optimization of these assignments is often done through trial and error method for achieving optimum performance of the FLC.

We have just only one output, node’s priority level, and have assigned fuzzy memberships as we did for inputs. It is shown in Table 3. The Fuzzy Inference Technique available of MATLAB is used in our node election method. The so-called mamdani method is applied. Fuzzy membership functions for input and output, fuzzy rule base and the defuzzification is shown in Figure 4. and 5.

The output which is node’s priority level for different inputs, i.e., none’s residual energy level, distance from the BS and trust levels is shown in Figure 6. Defuzzification of node’s priority level output is evaluated using the centroid approach: overlap and additive composition.

TABLE 1. Universe of Discourse for inputs and outputs.

Name Input/ Output

Min. Value

Max. Value

Node’s Trust (TL) I 0.0 1.0 Node’s Distance Level from BS (DL) I 0 10 Node’s Residual Energy Level (EL) I 0 1.0 Node’s Priority Level for Routing (PL) O 0 10

TABLE 2. Fuzzy variable ranges for different inputs.

I. Trust levels Crisp Input Range Fuzzy variable

0.0 – 0.30 SmallT 0.10 – 0.85 MedT 0.70 – 1.0 HighT

II. Residual Energy Levels Crisp Input Range Fuzzy variable

0 – 0.20 VWeakE 0.10 – 0.40 WeakE 0.30 – 0.60 MedE 0.50 – 0.80 HighE 0.70 – 1.0 VHighE

III. Node’s Distance from Sink Crisp Input Range Fuzzy variable

0 – 3 NearD 2 – 7 MedD 5 – 10 HighD

TABLE 3. Fuzzy Variable Ranges for node’s Priority Level for Routing.

Crisp Input Range Fuzzy variable 0 – 1.5 PL1 1.0 – 3.0 PL2 2.0 – 4.0 PL3 3.0 – 5.0 PL4 4.0 – 6.0 PL5 5.0 – 7.0 PL6 6.0 – 8.0 PL7 7.0 – 9.0 PL8 8.5 – 10.0 PL9

Fig. 4 Fuzzy membership functions for Inputs and Output

1282

b) Performance Evaluation To evaluate the performance of our proposed method, we have assumed one scenario of Wireless Sensor Network as shown in Figure 7. The node S is a source node and it is generating number of packets sequentially to transfer to the BS via neighbor nodes. And it has many neighbor nodes in its radio range, out of which, nodes A, B and C are main concerned. Here, node S has to select one best node among A, B, C to forward the packets to the BS. As we can see from the figure, nodes A, B, C are not only the neighbors of node A, but also they are neighbors of many other nodes of the network. Suppose, for transferring all the packet generated by node S, if only one node (node A or node B or node C) selected by the node S, may be due to its higher residual energy level or most trustworthiness or its smallest distance from BS, then that node’s (A or B or C) energy will be consumed all the time and its residual energy fall down to the threshold energy and may die very fast. Hence, this will definitely cause the total network to fail or at least some routes to fail from other nodes to the BS.

Suppose, if our method of node selection for routing is applied in the above situation, any one of the three nodes (A or B or C) will be selected at any given time based highest priority level given by FLC. That is the node selected for routing at time t0 may be different from the node selected at time t1, and so on. Also, the life span of the nodes as well as the entire network will be increased. In Figure 8, we have shown the graphical representation of the nodes selection by the FLC. The initial assumed values of trust, energy and distances for node A, B and C are shown in Table 4. From the figure it is clear that all three nodes are being selected based on their priority level given by the FLC. We can see from the figure that, all three nodes energies are being consumed smoothly by the source node S. For the shown 500 packets at different time instants, node A is selected for

Fig. 5 Fuzzy Rules

Fig. 6 Fuzzy output Routing Priority Level for different Energy levels, Distances and Trust levels

BS

S

A

B

C

Fig. 7 WSN lifespan test scenario.

1283

318 times, node B is selected for 136 times and node C is for 46 times. As shown in Figure 8, up to time T1 only node A was selected for routing, after that nodes A and B both are being selected up to time T2, and after that all three nodes A, B, C are being selected. Also, from the Figure 8, we can see that if the proposed routing method is not adopted, node A will die after the transmission of around 300 packets.

TABLE 4 Initial assumed levels Energy in

Joules Trust

(0.0 to 1.0) Distance (1 to 10)

node A 5 0.3 4 node B 4 0.4 5 node C 3.5 0.45 7

V. CONCLUSIONS The fuzzy logic based way of election of one node for routing among many neighbor nodes based on the trust level (TL), distance from the BS (DL) and residual energy (EL) presented in this paper provides the priority levels (Pi, i = 1 to 9) of all neighbor nodes for participating in the routing, and the highest priority level node may be selected for routing. We have shown graphically in this paper, how smoothly the priority levels are formed among the neighbor nodes, depending on the nodes trust level (TL), distance from the BS (DL) and residual energy (EL). Also, for an assumed scenario, we have shown the performance evaluation, i.e. how the lifespan of a node as well as the entire network are increased with this method. In future, we plan to develop a new routing protocol based on this new method of node selection for routing, and compare the efficiency in terms of energy consumption, memory requirement, processing cost and communication bandwidth comparing with different routing protocols available in literature.

REFERENCES [1] D. Culler, D. Estrin, M. Srivastava: Overview of Sensor Networks,

IEEE Computer Society, August 2004. [2] Chintalapudi K., Fu T., Paek J., Kothari N., Rangwala S., Caffrey J.,

Govindan R., Johnson E., Masri S., “Monotoring civil structures with a wireless sensor network”, Inernet Computing, IEEE, vol.10 no.2, pp. 26-34, March-April 2006.

[3] Mainwaring et al, “Wireless Sensor Networks for Habitat Monotoring”, International Workshop on Wireless Sensor Networks and Applications (ACM), Sep. 2002.

[4] C. Karlof and D.Wagner, “Secure Routing in Sensor Networks: Attacks and Countermeasures”, in First IEEE International Workshop on Sensor Network Protocols and Applications, 2003.

[5] A. Perrig, R. Zewczyk, V. Wen, D. Culler and D. Tygar, “SPINS: Secirity Protocols for Sensor Networks”, Wireless Networks, vol. 8, pp. 521-534,2002.

[6] Riaz A. Shaikh, Sungyoung Lee, M. A. U. Khan and Young Jae Song, “LSec: Lightweight Security Protocol for Distributed Wireless Sensor Network”, in Proc. of 11th IFIP International Conference on Personal Wireless Communication (PWC’06), Spain, Sep 2006.

[7] Chris Karlof, Naveen Sastry, and David Wagner, “TinySec: a link layer security architecture for wireless sensor networks”, in Proc. of the 2nd international conference on Embedded networked sensor systems, Baltimore, MD, USA, Nov 2004, pp. 162-175.

[8] Theodore Zahariadis, Helen C. Leigou, Panagiotis Trakadas and Stamatis Voliotis, “Mobile Networks: Trust management in wireless sensor networks”, EUROPEAN TRANSACTIONS ON TELECOMMUNICATIONS, vol.21, pp. 386-395, 2010.

[9] Yan Sun, Zhu Han, and K. J. Ray Liu, “Defense of Trust Management Vulnerabilities in Distributed Networks”, IEEE Communications Magazine, Feature Topic on Security in Mobile Ad Hoc and Sensor Networks, vol. 46, no. 2, pp.112-119, February 2008.

[10] Mayank Saraogi; Security in Wireless Sensor Networks; Department of Computer Science, University of Tennessee, Knoxville.

[email protected] [11] Javier Lopez, Rodrigo Roman, Isaac Agudo, Carmen Fernandez-

Gago, “Trust management systems for wireless sensor networks: Best practices”, Computer Communications, Elsevier, 33(2010) pp. 1086 – 1093.

[12] Shaik Sahil Babu, Arnab Raha, Mrinal Kanti Naskar, “Geometric Mean based Trust Management System for WSNs (GMTMS)”, World Congress on Information and Communication Technologies (WICT-2011), Mumbai, India, December 11-14, 2011.pp 444-449.

[13] Wendi Beth Heinzelman, “Application-Specific Architecture for wireless Networks”, PhD. Thesis, Massachusetts Institute of Technology, June, 2000.

Fig. 8 Indispensible node selection among three nodes

1284