An Efficient Routing Protocol for Secured Communication in Cognitive Radio Sensor Networks
aa r X i v : . [ c s . N I] A ug An Efficient Routing Protocol for SecuredCommunication in Cognitive Radio SensorNetworks
Sharmin Akter , Mohammad Shahriar Rahman and Nafees Mansoor Department of Computer Science and Engineering, University of Liberal Arts BangladeshEmail:[email protected] , [email protected] , [email protected] Abstract —This paper introduces an efficient reactive routingprotocol considering the mobility and the reliability of a nodein Cognitive Radio Sensor Networks (CRSNs). The proposedprotocol accommodates the dynamic behavior of the spectrumavailability and selects a stable transmission path from a sourcenode to the destination. Outlined as a weighted graph problem,the proposed protocol measures the weight for an edge themeasuring the mobility patterns of the nodes and channelavailability. Furthermore, the mobility pattern of a node isdefined in the proposed routing protocol from the viewpointof distance, speed, direction, and node’s reliability. Besides, thespectrum awareness in the proposed protocol is measured overthe number of shared common channels and the channel quality.It is anticipated that the proposed protocol shows efficient routingperformance by selecting stable and secured paths from source todestination. Simulation is carried out to assess the performanceof the protocol where it is witnessed that the proposed routingprotocol outperforms existing ones.
Index Terms —Cognitive Radio Sensor Networks; V2V Com-munications; Ad Hoc Networks; WSN; Routing Protocol
I. I
NTRODUCTION
With latest advancement of the wireless communication,Wireless Sensor Networks (WSN) are the devices consistingof large number of low powered battery sensors which has theability to sense, transmit data and relay to other sink nodes.It has earned profound popularity in wide range of sensorbased application in different wireless communication areas.It is designed for fixed and licensee free Spectrum. Hence,there results in excessive interference occur during sharingsame channel in data dissemination. As a result, packet lossor network failure is endangered due to this interference. Soefficient utilization of spectrum has resulted in a challenge.To overcome this challenge, CR technology has emergedas a prominent technology which utilize the radio spectrumefficiently in a opportunistic manner [1] and thus improvesspectrum efficiency. By providing CRN technology to WSNwhich prompts to Cognitive Radio Sensor Network (CRSN),WSN nodes capable to get the facilities of the CRNs and willbe able to exploit the utilized spectrum intelligently. In this paper a new trust based secure routing mechanismis introduced for route discovery and selection process con-sidering spectrum scarcity and dynamic mobility pattern ofthe spectrum band. Here, mobility, distance, link stability areconsidered both in spectrum sensing and data transmission[14]. A clustering mechanism is also presented in this paperfor discovering stable route for faster data delivery.Many protocols of CRSNs are proposed in the previousresearch work. Based on the limited energy of the routingprotocol has been proposed in [6] but it does not consider spec-trum mobility and lacks secure communication. For assuringsecure communication detection of malicious node algorithmis proposed in [11] which improves network performance con-sidering distrust value. A cluster based scheduling algorithmhas been proposed in [4] which considers sensor energy andthe collision of the CR nodes in the clusters. Based on theenergy consumption in CR nodes, a cluster based routingis demonstrated in [5] which reduces energy consumptionand also improves free channel utilization. By implementingdijkstra algorithm an energy efficiency routing protocol in[8] which is optimized to avoid interference of PU user inchannel switching. Considering mobility, distance, velocity ofCR nodes routing protocol has been proposed in [10]. Then,based on velocity and location of the nodes the presentedrouting protocol [7] does not consider spectrum availability,link quality and reliability. Based on the location and density,a software based routing protocol is presented in [9] whichalso does not utilize the spectrum bands and does not discussabout the link quality, collision of CR nodes and reliability.The rest of the article is organized as follows. Section IIdepicts a network model of the proposed routing protocol.Section III discusses the proposed protocol and the routingmechanism. Then simulation results and performance evalua-tion are discussed in section IV. Finally, the paper concludeswith the conclusion and future work in section V.
I. N
ETWORK M ODEL
In this proposed routing protocol, each CR in the networkhas the ability to sense the free spectrum bands associated withtwo transceivers equipment for control and data transmission.It is assumed that the transmission range is equivalent toall nodes and each CR node is capable to compute theNHDF (Next-hop Determination Factor) [14] and IF (IntrusionDetection Factor). To distinguish the licensed spectrum ofPU’s, SU observes its nearby radio transmission. The proposedspectrum aware clustering mechanism separates the networktransmission system into legitimate groups. By propagating theneighbour nodes’ list after the neighbor discovery is executed,ACL exchange is accomplished among 1 -hop neighbors. Todiminish the number of channels and furthermore develop-ment of clusters with maximum common channels, a clusterformation mechanism, earlier published by the author, namedRARE is utilized in the proposed routing protocol [13].Here, both CR nodes anticipate link quality for estimatingthe stability of the transmission between two nodes in thenetwork. The source node chooses the following neighboringnode with a minimum delay which has better link quality.Here, each node knows about traffic density as having a similarchannel by the node causes congestion. For increasing the se-curity a concept of blockchain consensus mechanism flowchartis proposed in Fig. 1 for detecting any abnormal behaviorof vehicles or any internal or external attack for example,DoS, Black-hole, sticking, and so on which keep track of thetrust level and forwards safe messages for transmission thoughdiscarding the malicious node to achieve stable and reliablecommunication in the same network transmission.III. P
ROPOSED R OUTING P ROTOCOL
In this routing protocol, a routing metric namely
N HDF is presented [14]. Here, total cumulative N is determined withrespect to routing path which is estimated as follows, N i,j = ( ξ T δ Ei,j ) C n IF (1)Here, ξ T is transmitting weight value dependent on trans-mission range ( Φ t ), displacement τ v , link delay δ Lp and speed(s). ξ T = Φ t τ v ∗ δ LP ∗ s (2)At the time T the displacement ( τ v ) between two neighborvehicular nodes is calculated below, τ v = dθ (3)At first during broadcasting RREQ distance ( d ) is estimatedmeasuring the RSSI value which is demonstrated as follows, d = 10 ( κ − log ( 4 πl υ )10 ω ) l (4) Here, where k is considered as path loss, ω is the signalloss exponent, υ is the wavelength of the received signal and l is reference distance.During this data dissemination the sensor node’s speed (s)can be expressed as follows, s = p ( α r − γ r ) + ( α t − γ t ) ( T + ∆) − T (5)where, T and T is the received and sending timestamprespectively alongside present co-ordinates by waiting for arandom back-off time. Here, ψ r ( α r , γ r ) is denoted as the co-ordinate of sensor nodes during receiving and ψ s ( α t , γ t ) ischaracterized as the co-ordinate of the vehicular node duringsending. ∆ is denoted as transmission time. Then, directionbetween neighbor node and the destination node is estimatedas follows, θ = cos − ( −−−→ ψ r ψ s −−−→ ϕ r ϕ s ||−−−→ ψ r ψ s || × ||−−−→ ϕ r ϕ s || ) (6)where, ϕ r ( α s , γ s ) is the destination position during receiv-ing data packets and ϕ s ( α q , γ q ) is while sending data packets.Three kinds of delays such as back-off delay, queuingdelay and switching delay are considered in this routingprotocol between the the intermediate nodes. If the quantityof neighboring sensor node is resolved as V i , RT i is indicatedas data rate of N i and S is the size of the data, the queuingdelay ( δ NL ) of N i can be evaluated as follows, δ Ni = SV i RT i (7)Sensor nodes utilize random back-off time as various sensornodes may utilize a similar channel to avoid collision duringsame channel sharing. By the below equation the back-offdelay can be estimated, δ Ki = 1(1 − b c )(1 − (1 − b c ) V i − ) z (8)here, b c is recognized as the probability of collision and V i is the neighboring sensor node on a channel CH i and z is thewindow size.Hence, during forwarding the data packets if a node N i is required to switch from channel p to divert q in itschannel-group during sending the message to next-hop N j , thecalculated channel switching delay is characterized as follows, δ Mi,j = a ∗ | p − q | (9)where a is positive real number and for specific step size a is considered as the tuning delay of two neighboring channels.For step size 10MHz, a is denoted as 10ms [14].Thus, the delay of the link ( δ Ei,j ) that connects N i and N j is calculated using Equation 7, 8 and 9 as below, δ Ei,j = δ Mi,j + δ Ni + δ Ki (10)he Intruder Determination Factor ( IF CN j CV i ) of the neigh-boring node ( CN j ) for any sensor node ( CV i ) can be evaluatedas follows, IF CN j CV i = e RN (11)here, RN indicates the report number during each questionor suspect by vehicular nodes during transmission. TABLE IS
YMBOLS USED IN THE PROPOSED ROUTING PROTOCOL
Symbols Description CH Cluster head CS i Neighbors of channel set V j Joining node
RREQ
Route Request during data transmission CN j Neighbors of cluster set CV i Any sensor node in cluster
MCV i Malicious behavior of any vehicle CV i IF Intrusion detection factor QM s
50% malicious suspect or query by any CV i RN Report number BV i Behaviour of CV i A. Route Discovery and Selection Process
In the proposed routing protocol, at first the joining node ofa cluster V j starts spectrum sensing. Then the V j broadcastsaccessible channel set CS i to every neighbor of cluster set CN j . Then RREQ is broadcast by CN j While broadcastingRREQ the neighboring or relay nodes observe the behavior of CV i and then estimate IF value by implementing intruderdetection mechanism demonstrated in Fig. 1. Then CN j updates the routing table with IF value. When the destinationnode exists in the neighboring list, the neighboring nodeestimates the N DHF value by equation 1. After that clusterhead CH initiates a RREP with N HDF and IF with therouting path to the previous generated RREQ message. Thisis a continuous process until the neighbor node receives theRREP message and all possible routing paths are discovered.Lastly, the N HDF and IF values are stored in in routingpath array calculated by the source node. B. Reliable Data Dissemination Mechanism
Here, in the intruder detection mechanism every nodejoins in the blockchain network and gathers information fromthe neighboring sensor nodes. During sending RREQ everyneighboring vehicle CN j observes the behavior of any sensornode CV i in the blockchain network. When the neighboringnodes find any report or query or suspect of showing abnormalbehavior of V i , IF CN j CV i increments where the value of IF CN j CV i is initially is set to 1 and report to the all neighboring nodesin the blockchain network. Then the neighboring node updatesthe routing table with new IF CN j CV i value. CN j checks if morethan query or suspect of malicious behavior exist, the IF CN j CV i is set as INFINITY which detects CV i as maliciousnode and sends warning message to the all the neighboringnodes in the network. Finally, the entry of node CV i is updatedin blacklist and the malicious node CV i is discarded from the Fig. 1. Flowchart of the Proposed Routing Protocol network. Thus consequently we can monitor all the vehicles’present trustworthiness.IV. S
IMULATION R ESULTS
The performance of the proposed routing protocol is eval-uated by conducting simulation on NS2 which is a discreteevent simulator. The network setup is associated with 10clusters consisting of 400 sensor nodes. PUs and SUs arerandomized position in this network where other configura-tions are listed in the Table II. To analyze the performancea comparison study has been conducted with the existing umber of Nodes
10 30 50 70 100 A v e r age E nd - t o - E nd D e l a y ProposedDMNPRACBOCAS a) Comparison in terms of End-to-End Delay
Number of Nodes
10 30 50 70 100 T h r oughpu t ProposedDMNPRACBOCAS b) Comparison in terms of ThroughputFig. 2. Performance comparisons of the proposed routing protocol with other protocols. routing protocols namely OCAS [1], PRACB [6] and DMN[11] considering throughput and end-to-end delay.
TABLE IIS
IMULATION E NVIRONMENTS
Parameters Value
Number of vehicular nodes 10,30,50,70,100Simulation area 4000 m Run time 120 sPacket size 256 bytesTraffic type CBRCBR 5 packets/sQueue type Drop-tailTransmission range 500 mPropagation model Two-Ray GroundMac Layer Protocol 802.15.6pNumber of malicious nodes 5speed of vehicles 5 m/s
Fig. 1(a) shows the comparing results of the proposedrouting protocol with other existing routing protocols OCAS,PRACB and DMN in terms of End-to-End delay. Here, thedelay increases as the density of the network increases withthe number of nodes and this requires more time which resultsin more delay. Here, PRACB performs better as it lacksof considering malicious or any abnormal behavior of nodeduring data transmission. On the other hand, the proposedrouting protocol which shows little bit higher delay rather thanPRACB, considers reliability and observes malicious behaviorfor intruder detection. That is why the trusted and reliablepath selected by proposed routing protocol is longer than theshortest one during data transmission.Fig. 1(b) depicts that our proposed routing protocol outper-forms than other existing routing protocol OCAS, PRACB andDMN. Here, throughput increases with the number of nodesas more sensor nodes are connected in the network for packettransmission which causes less packet drop or network failure.The proposed routing protocol performs better as it considersmobility as routing metric and also estimates stable route forboth in spectrum sensing and data transmission. V. C
ONCLUSION AND F UTURE W ORKS
A secured reactive routing protocol for CRSN is presentedin this paper. The double-folded dynamic behavior of thenetwork is considered in the proposed protocol where au-tonomous movements of the sensors and changing spectrumavailability are measured. From the simulation results, it isobserved that the proposed routing protocol performs betterthan other recently introduced protocols. This study will leadto the further development of the routing protocol where theenergy efficiency of the network will be addressed.R