15-09-2016, 03:12 PM
1454660770-projectproposedpaper.docx (Size: 38.33 KB / Downloads: 4)
ABSTRACT
Vehicular ad hoc network are created by applying the principles of mobile ad hoc networks. The instantaneous creation of a wireless network for the data exchange to the domain of vehicles is the key factor of vanet.
VANETs enable wireless communication among vehicles and vehicle to infrastructure. Main objective of VANET is secure, comfort and convenience on the road. Unlike simple ad-hoc network VANET differs by its unique characteristics. However, because there is no centralized administration, it becomes vulnerable to misbehaviors. This in turns threatens different aspects of VANET being such a useful network must provide adequate security measures for secure communication. The proposed MHRV algorithmin VANETs improves various algorithm such as DMV in terms ofeffective selection of verifiers for encountering of malicious nodes and hence improves the network performance as well as flexibility.
MRHV: MALICIOUS NODE REMOVAL HEURISTIC IN VANET
1. INTRDUCTION
The communication techniques have improved to its higher level, especially in signal processing and microelectronics. Low consumption of energy sensorial capacities and wireless communication are the fields where communication techniques have evolved most.The organization of sensory devices micron in one sophisticated computational architecture and of communication allowed creating a new type of net without wire.
This new technology allows to monitoring remote environment and to get more required information, taking off advantage of being employee well next to the phenomenon of interest. This new technology with diversity and glimpsed potential has motivated considerably the research in the area. The wireless sensory applications are extended for diverse areas of interest as, for example, in the medicine, the ambient monitoring, transportation, installations security, amongst other possible ones.
VANET (Vehicle Ad-hoc Networks) is emergingtechnologies that they deserve the attention of the industry and the academic institutions. The vehicular communications (VC) meet in the center of numerous initiatives of the research that enhance the security and the efficiency of transportation systems, supplying, for example,report of the ambient conditions snow, fire, etc., conditions in the road such as emergency, construction sites, or congestion.
VANET is a wireless communication network in which moving objects such as vehicles are connected instead of computers. VANETs enable wireless communication among vehicles and vehicle to infrastructure. Its main objective is to render safety, comfort and convenience on the road. However, because of lack of centralized administration, it becomes sensitive to misbehaviors. This greatly threatens different aspects of VANET being such a useful network must provide proper security measures for safe communication. The proposed algorithm MHRV-Malicious Node Removal Heuristic in VANETs improves Encounterion of Malicious Node inVanet algorithm in terms of verifierselection in an effective way for encounterion of malicious nodes. The MHRV algorithm selects all the nodes as verifiers which have distrust value less than the vehicle to be monitored.
2. LITERATURE REVIEW
Number of schemes has been proposed to encounter misbehavior and malicious nodes in Vehicular Ad-hoc Networks. The misbehavior encounterion schemes can be broadly classified into following two types: Node-centric and Data-centric misbehavior encounterion schemes.
2.1. Node Centric Misbehavior Encounterion Schemes
Node-centric technique distinguishes different nodes based on authentication. Security credentials, digital signatures, etc are used to authenticate the node which transfer the messages. This type of schemes give importance to the node which transmit the messages rather than the data transferred.
In the research work 9, 10, Gosh et al. have proposed a robust scheme to encounter malicious vehicles for Post Crash Notification application. They have considered the probability of faking the information of position of the vehicle in the PCN along with the false crash alert in 10. Kim et al.6 have proposed a novel Misbehavior Based Reputation Management Scheme (MBRMS) which includes three components a) Misbehavior encounterion b) Event rebroadcast and c) Global eviction algorithms for the encounterion and filtration of false information in vehicular ad-hoc networks. Daeinabi et al.3 have proposed a encounterion algorithm called DMV to discover malicious nodes through observations that duplicates or drops received packets and isolates such vehicles from honest nodes. Vehicles are tagged using a distrust value and are monitored by the allocated verifier nodes. Wahab et al.16 have used Quality of Service-Optimized Link State Routing (QoS-OLSR) clustering algorithm to encounter malicious vehicles in (VANET) using Dempster±Shafer based cooperative watchdog model. This method maintains stability and quality of service with increase in encounterion probability and decreases the number of selfish nodes and false negatives. Kadam et al. have presented a new approach14 for not solely the encounterion of malicious vehicles attack, additionally their prevention from the VANET. It is an improvement of the Encounterion of Malicious Vehicles (DMV) algorithm 3. This approach reduced the impact of black hole attack within VANET and is more efficient and secure compared to DMV.
2.2 Data Centric Misbehavior Encounterion Schemes
The method of Data-centricchecks the transmitted data among nodes to encounter misbehaviors. It is primarily concerned with a linkage between the messages than the identities of the individual nodes. The information propagated by the nodes in the network is evaluated and correlated with the information received by the other nodes, in order toverify the authenticity about the alert messages received. A minute number of research contributions to the data-centric misbehavior encounterion scheme are as below.
In the research work2, Vulimiri et al. have encountered misbehavior in VANETs based on the secondary alerts or information that are created in return to the primary alerts for PCN application. A new misbehavior encounterion scheme that is introduced by Ruj et al.20 uses the concept of data-centric misbehavior encounterion algorithmic program. Fake alert messages and nodes that are misbehaving are encountered by monitoring the actions of the vehicle after alert information have been sent. To be able to find the vehicle showing malevolent behavior, Rezgui et al.18developed a mechanism VARM that collects, at one vehicle, information each and to every neighbor’s transmission. Rawat et al.17 have suggested a novel algorithm to secure the communication in the VANET by encountering malevolent driver using a probabilistic approach. It computes the trust of the received messages and analyses whether the message is from an honest vehicle or not. Grover et al.12 have presented a security scheme in order to categorize numerous misbehaviors in VANET using machine learning technique. It comprehends a malicious node and an honest node based on the features computed by the observer nodes. Grover et al.13 presented a security scheme for encounterion of misbehaviors in VANETs using an ensemble primarily based machine learning approach. Based on systems to encounter misbehavior, running on vehicles and roadside infrastructure units, a central evaluation system15 is introduced that aims to find and exclude attackers from the network. In the research paper4, Barnwal et al. have introduced a brief term misbehavior encounterion framework which can encounter a malicious node that is spreading fake location and speed information along its heartbeat/beacon messages. Harit et al.11 have presented a scheme based on the data centric concept which encounters the authenticity of the information received, primarily for PCN alerts. It makes utilises a Fox-Hole area which helps to find the safety value of any vehicular node on its current location and present speed. In the paper 7, Huang et al. have suggested a cheater encounterion protocol which encounters malicious vehicles that newscast fake crowding information in the network for their selfish motives and imitate other non-existing vehicle. This concept is mainly based on dimensions of local distance and velocity by means of radars to certify the crowding event sent by a vehicular node. Coussement et al. 19 have suggested an Intrusion encounterion system (IDS) which has the ability of encountering malicious actions made to the system. A decision support agreement is prepared for security in VANETs which authenticates the signature of the incoming and outgoing packets.
3. NETWORK MODEL
Two main components of VANET are vehicles and Road Side Units (RSUs). The communication is using short range radio communication. An authority of VANETs called Certificate Authorities (CAs) are the one who is responsible for providing security and authentication. The main component responsible for management of identities of the vehicles in the network and verifying the misbehavior is certificate authority (CA), the reports sent by the verifier nodes is collected by certificate authority and if found true, the distrust value of nodes are modified accordingly. There are two types of lists stored by vehicle black list containing list of malicious node sent by CA and white list provided by its respective cluster head or verifier.
The characteristics of a vehicular ad hoc network are unique compared to other mobile ad hoc networks. The distinguishing property of a VANET offers opportunities to exploit these properties to increase network performance, and at the same time it presents considerable challenges. The characteristics of a vanet network are given below.
• First, a VANET is characterized by a rapidly changing but somewhat predictable topology
• Second, frequent fragmentation of the network occurs.
• Third, the effective network diameter of a VANET is relatively small.
• Fourth, redundancy is limited both temporally and functionally. Last, a VANET poses a number of unique security challenges
3.1 ALGORITHM DESCRIPTION
The algorithm is a node centric misbehavior encounterion scheme. In Node-centric techniques different nodes are distinguished using authentication. Digital signatures, security credential etc are used to authenticate the node transferring the message. Such schemes emphasis on the nodes transmitting the messages rather than the data transferred The Encounterion of Malicious Node algorithm is based on the following three basic concepts.
1.A vehicle is said to be malicious if.
a) If it drops or duplicate packets received.
b) If it create congestion in the network.
c) If it misguide other vehicles nodes or destroy crucial information.
2. An honest vehicle forwards the messages received to it correctly to other nodes in the network or creates right messages for transmission.
3. If the vehicle repeats to show abnormal behavior such that if its Distrust value exceeds the threshold value TMD then vehicle will be tagged as a malicious vehicle.
3.2 DESIGN
In VANET communication, a node acts as a source of information. There is a second type node which acts as a destination of the message. The third kind of node is intermediate nodes between source and destination acts as relay nodes. When a node say VN plays the role of a relaying node, other honest vehicles, which monitors its behavior. When vehicle VU works as a verifier of VN, it checks the number of messages received by VN (represented by parameter a) and number of messages that VN drops or duplicates as encountered by VU (represented by parameter b). After a particular time has expired say PL, if vehicle VN does not forward a received packet or sends its numerous copies, it is considered as abnormal behavior by verifier VU. Then the verifier increases the value of parameter b by 1 unit. The parameter DV (distrust value) is associated with each vehicle and changes when an unusual behavior is observed. The new distrust value is broadcasted to all neighbors and they update their lists accordingly. Vehicles cooperate with one another while they are part of the white list as their DV is lower than the threshold. If it exceeds the threshold, the ID of the vehicle is reported to the CA as a malicious node. CA then broadcasts the identity of malicious node to all others nodes in the network. In the proposed algorithm, verifier is selected on the basis of the parameters: distrust value, load, power ratio and distance. The node is selected as verifier when Decision parameter, DP is less than the Selection Threshold, TVS among other neighboring nodes located in the region r ( VN). The proposed approach optimizes the selection of verifier by considering different cases. It helps to save the network bandwidth and hence improves network performance. Vehicular nodes in the locality r are considered for being verifiers. The region r is calculated as the intersection area of transmission rage of VN and its CH. Area of CH means its transmission range and area of vehicle VN is calculated by the formula given below in Eq (1). Thus it ensures all verifiers are able to send misbehavior reports to the CH.
Area (VN) = TR(VN) - PL (Smx - Smn ) (1)
Where,
TR (VN) - Transmission range of vehicle VN.
PL - Packet latency in vehicles.
SMX - Maximum speed of vehicle.
SMN - Minimum speed of vehicle.
The parameters for selection of verifiers in the area r are explained below.
1. Load (LD) – The load means total number of vehicles that a node is currently monitoring. It is considered so as to equity the monitoring job among different nodes. A node which is having less load compared to other nodes in the network will have greater chance of being as verifier.
2. Distrust value (DV) - It refers to the measure of trustfulness of a vehicle. It means the node is trustful when the distrust value is less. If a vehicle shows strange behavior, this value is increased and compared to the threshold for making appropriate decisions, ie a vehicle should remain in the white list or tagged as a malicious vehicle and moved to the black list.
3. Distance (DS) -If the distance of a node from the vehicle to be monitored is less, then the node will remain in the transmission range of the vehicle for a longer time. Thus, this provides scope for better observations and decision making.
4. power to weight ratio(PR)- power ratio represents the ratio is a measurement of actual performance of any engine or power source
Decision Parameter, DP is calculated for all the nodes considering for verifier selection by considering the parameters load, distance distrust value and power to weight ratio of the node by the following equation (2).
DP = W1 * LD + W2 * DV + W3 * DS+ W4 * PR(2)
where,
W1, W2, and W3 W4are the weight factors for parameters Load (LD), Distrust Value (DV, Distance (DS) and Power ratio (PR) respectively
such that,
W1 + W2 + W3 +W4= 1 (3)
The equation is further divided into 2 cases.
Case 1: Heavy traffic situation. W1=0
Case1 deals with the heavy traffic situation. More number of vehicles should be verified by each verifier during heavy traffic scenario which implies value of w1 should be almost zero. That is we don’t consider the term containing load this in turn improves the flexibility as well as time complexity of the equation, hence equation becomes
DP = W2 * DV + W3 * DS+ W4 * PR
Case 2: For light vehicles. W4=0
Case 2 deals with the power to weight ratio. Power to weight ratio is more considerable for heavy vehicles such as goods carrying vehicles. Hence we take w4 as zero for light vehicles with the assumption that they maintain a good power to weight ratio. Hence the main equation becomes
DP = W1*LD+W2 * DV + W3 * DS
This in turn improves the flexibility of the equation.
Rather selecting all the nodes with smaller distrust value than the vehicular node VN, allocating few nodes as verifiers which are more suitable for monitoring process helps in improved encounterion of malicious nodes as well as improves performance of network. As few nodes perform the job of monitoring the node VN, this saves network resources used for reporting the behavior and conserve their time for processing the observed behavior for all the nodes. As the network utilization is improved, it results in better transmissions in the network. In order to assign verifiers for the node VN, decision parameter DP calculated for the nodes under consideration is compared to the Selection Threshold TVS. If a nodes decision parameter value is less than the selection threshold (DP < TVS), then the vehicle is allocated as verifier. This way proposed approach optimize the selection of verifier nodes. Vehicles know the distrust value of other vehicles present in its neighborhood. When a vehicle VU reports a misbehavior of another vehicle VN, the CH verifies the DV of VU to confirm that it is equal or lower to the DV of VN. CH is considered to be the most reliable and trustworthy node within a cluster. Thus, verifiers for an honest node are not assigned among the vehicles which show abnormal behavior as such vehicles have greater DV as compared to a normal node. In case, CH is found to misbehave then it is replaced by a trustier vehicle. Hence, the process consists of all aspects of monitoring the vehicles in order to identify the malicious nodes. It also improves the selection of verifiers by the proposed approach which results in better network utilization and enhanced performance.
Selection Threshold:
In practical implementation of the network we choose a constant value for selection threshold. By considering the factors such, as node density, channel quality, energy of node, signal strength and etc.