Enhanced AODV Routing Protocol for An architecture ...

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/

Enhanced AODV Routing Protocol for An architecture V2V in VANET Dr Hassan NAANANI1 ,Youness Farah2, Mohamed Ouamer3, Ibtihal Mouhib4 1Faculty

Of Sciences Ben-msik university Hassan II Casablanca, Morocco; of Sciences Ain Chock, Hassan II University of Casablanca, Morocco; 3Centre Des Classe Prépa Ecole Okba Dakhla, Morocco 4Moulay Smail University, Faculty of sciences, TAR Lab Meknes, Morocco; 2Faculty

Abstract Direct communication between vehicles has become a useful and effective in the management of Road Traffic. It is therefore necessary to implement the multi-hop technique to convey information. From this is born intelligent transport systems (ITS), which aims to improve efficiency and passenger comfort. This architecture is known dime name VANET (ad hoc network Vehicle). There are two categories of wireless networks, network infrastructure with V2I which generally use the cellular communication model, and without V2V infrastructure named ad hoc networks. In this paper, an improved version of AODV protocol is developed to refine the management of the saturation of the queue for nodes in a VANET network. This improvement is achieved by optimizing the packet residence time within the queue of each node. Simulation results show that the improved AODV protocol.

Keywords: VANET, sumo, AODV, V2V

1. Introduction: Currently, every year many people around the world die due to vehicle accidents, causing a security issue, such as traffic and speed limits lights are used, but it lacks a mesh solution for this issue. Governments and number of motor vehicle companies also admitted that the road traffic security is very difficult. So, to improve the safety of the vehicle and road users, a new technology was emerged named VANET[5]. VANETs are an refined type of MANETs. A VANET network consists of vehicles,which are considered as mobile nodes to build up a robust ad hoc network. VANET is built based on an inter-vehicle communication. Moreover, apart from accidental and safety features, there are also wide ranges of applications that exist in VANET, especially those that can provide passenger comfort such as web browsing, predictable mobility through updates GPS, and so on. VANET networks are the new spreading version of Mobile Ad-hoc Network (MANET) where the mobile nodes consist only of cars, buses, nd other types of vehicles.[1] Recently, vehicular ad hoc network becomes a new research topic due to its independent nature of the network infrastructure such as base stations. The nature of the infrastructure of the least dynamic VANET is and reapplies series of networks to implement strategies to ensure efficient and reliable communication from end to end between vehicles. VANETs are easy to manage and to deploy. In the future, automated packet routing between vehicles will be mandatory. In this paper, we will try to improve queue management in reactive routing protocols used in VANETs, and we will test our work by measuring the following characteristics: PDR (Packet Delivery Ratio) and NRL (Normalized Routing Load), E2E (End-to-End Delay) and Packet Loss Ratio (PLR). Our VANET network is simulated using NS3 [6] (network simulator). In order to have more realistic simulation results, we developed a mobility model that reflects the movement of vehicles in city scenarios using SUMO traffic simulator[12,14]. In this paper the real time traffic simulation created for Swindon magic roundabout area is shown in Fig2.

1.1 City Scenario

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/ Our city scenario is based on the traffic roundabout shown in Fig1. The Magic Roundabout in Swindon, England, and is made up of five mini-roundabouts and sixth central, anti-clockwise roundabout. In this paper, we developed an improved version of AODV protocol to refine the management of the saturation of the queue for nodes in a VANET network.

Fig1- Swindon magic roundabout

Fig2 - Map of Swindon Magic Roundabout imported into SUMO Traffic Simulator [8] 2. RELATEDWORK In ref, each queuing algorithm is efficient in some area and not in another one. For example. The research demonstrates that REM has the maximum PDR, while SFQ has the PDR. Additionally, RED can deliver packets with the lowest end-to-end delay. In ref, they choose to handle packets using queues with different lengths in order to decrease packet-dropping ratio and avoid the deficiency of long route establishment on nodes with fixed queue length.

3. OVERVIEW OF AODV • The AODV protocol [2] is essentially an improvement of the DSDV algorithm. The AODV protocol reduces the number of broadcasts of messages, and in that when creating the paths necessary, unlike the DSDV, which maintains all the paths.

• The principles AODV uses sequence numbers at the end to maintain the consistency of routing information.

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/ • Because of the mobility of nodes in ad hoc networks, roads change frequently so that the roads maintained by some nodes become invalid. The sequence numbers allow the use of the newest ways (fresh roads). • In the same way as in the DSR [7], the AODV uses a route request in order to create a path to a certain destination. However, the AODV [2] maintains a distributed paths maintaining a routing table at each gateway node belonging to the searched path way. • A node sends a (RREQ) route request in the case where he would need to know a path to a certain destination and such a route is not available. This can happen:  If the destination is not known in advance, or If the life of the existing path to the destination path has expired or is become faulty. • When the destination receives this message, it responds with a route request (RREP). • The RREQ destination field includes the last known value of the sequence number linked to the destination node. The sequence number is stored in the routing table. If the sequence number is not known, the node zero as a value for it. The sequence number of RREQ source is the same as in the source node. • In the last phase of maintaining consistent routes, periodic transmission of "HELLO" is performed. If three messages "HELLO" are not consecutively received from a neighbor node, the link in question is considered faulty. • The AODV protocol does not present a routing loop, it also avoids the problem "counting to infinity" Bellman-Ford, which provides fast convergence whenever the topology changes.

4. .PROPOSED SOLUTION Finite Buffers: M/M/1/ In order to decrease delay on a VANET network, we propose a model based on queuing theory. The principle of this model is as follows. A wireless node can be seen as a buffer, which is filled by incoming packets from higher layers. Additionally, a single server on the network is dedicated to handle these packets. We can model this by a M/M/1 waiting queue system as shown in Fig3, having the following properties:    

The arrival of packets follows an exponential distribution with lamda parameter. The treatment of incoming packets also follows an exponential distribution with parameter μ. There is a single server processing incoming packets. The delay of a packet in the queue of the transmitter simply equals the average time a customer rated R arriving in the queue, given by Little's formula[4,8]

4.1 Study of the stationary state  The property of exponential laws is not to have memory: the future state depends only on the present state, regardless of the past (Markov random system). Moreover, the number of items in the M/M/1 queue is enough to illustrate the state and progress of the system.  The average number of items in the system per time interval and the average number of items N being in the queue + server system can be calculated directly from the following probabilities pk:

 Differentiating each member See compared to ρ :

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/  By substituting See into the equation, we obtain the value of N:

 Note that when ρ approaches 1, N tends to infinity

5. PERFORMANCE EVALUATION In this section, the evaluation study is performed using the discrete-event simulator ns3 [[3]]. In the rest of this section, we will define simulation parameters, mobility scenarios, and performance metrics; and describe the road traffic simulated [9].

5.1. Simulation parameters and metrics In this work, we analyze the performance of new enhanced protocol AODV and AODV-AV routing protocols using NS3 (network simulator) and Sumotraffic simulator [6]. These software programs help simulate and examine VANET traffic and packet transmission between the nodes in the network. In this simulation, we test routing between a set of 20 and 80 nodes using AODV and AODV-AV protocols. Some of the parameters included in the experimental analysis arelisted in simulation table below.

5.2 Network Simulator A network simulator consists of a wide range of networking technologies and protocols and help users to build complex networks from basic building blocks like clusters of nodes and links. Table 1 -NS3 Simulation Parameters

Parameter Name Network simulator Network interface type Routingprotocol Interface queue type Queue length Time of simulation end Number of nodes in topography Area Node placement Traffic type Radio propagation model

Value NS3, Sumo Traffic Simulator Physical wireless AODV, AODV-AV Priority queue 50 packets 100 simulation seconds 20 and 80 81 X 81 and 163 X 163 Random TCP Two Ray Ground

The performance of the improved AODV is measured and tested against basic AODV under different node speeds and densities. The performance metrics that we've used are described in the next section.

5.3 Performance metrics Another important aspect to consider when evaluating routing protocols is which performance metrics should be used in order to represent an accurate performance of the routing protocols. It is vital to utilize metrics that

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/ demonstrate the performance of the routing protocols in a range of conditions. Below are some of the most used metrics in simulation-based studies intentended to measure the perfomance of a routing protocol. End to end delay (E2E): in a mobile adhoc network, there will be delays due to buffering during queuing at the node's interface level, route discovery, packet retransmission and broadcast, and transfer times. The average E2E as in [2] is the average time needed to transmit data packets from source node to destination node. To calculate the average E2E, we recorded the time difference between the instant CBR packet was sent at and the instant it was received at. Then, we divide the total calculated delivery time over the number of CBR packets received. This metric describes the packet delivery time. The lower the average E2E the better the protocol performs. E2E can be defined as the time need for data packets to be transmitted across an ad hoc network from the source to the destination node.

𝐸2𝐸 (𝑠) =

∑(Delivered Time − Transmitted Time) 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑃𝑎𝑐𝑘𝑒𝑡𝑠 𝑠𝑢𝑐𝑐𝑒𝑠𝑓𝑢𝑙𝑙𝑦 𝑑𝑒𝑙𝑖𝑣𝑒𝑟𝑒𝑑

Packet delivery fraction (PDF) is the ratio between total data packets received over total packets sent by sources as given in [1]. The PDF illustrates how successful a routing protocol delivers packets from a source node to a destination node. Higher PDF means better results. This metric distinguishes both the completeness and the correctness of the routing protocol.

𝑃𝐷𝐹(𝑘𝑏𝑝𝑠) =

∑ Delivered Application Packets ∑ Sent Packets

Normalized Routing Load (NRL): The number of routing packets transmitted per data packet delivered at the destination. Each hap-wise transmission of a routing packet is counted as one transmission.

𝑁𝑅𝐿 =

∑ Routing packets ∑ Delivered Application Packets

6. Simulation E2E

Fig1 E2E vs Vehicul Density, 2O Nodes


Fig2 E2E vs Vehicul Density, 8O Nodes

PLR

Fig3PLR vs Vehicul Density, 2O Nodes

Fig4 PLR vs Vehicule Density, 8O Nodes NRL


Fig 5 NRL vs Vehicul Density, 2O Nodes

Fig 6 NRL vs Vehicul Density, 8O Nodes

7. Conclusion In this paper, we evaluated the performance of two routing protocols used in mobile wireless networks (AODV and an improved version of it). We measured their packet delivery ratio, average end-to-end delay and normalized routing load in a situation simulating real road traffic of Swindon magic roundabout. We conclude that it is straightforward to say which of the protocols has better performance than the other. The results illustrate that AODV-AV (AODV Applied to VANETs) performs better than AODV according to the metrics listed above. Finally, the packet delivery ratio, the E2E and the normalized routing load are affected by vehicle density in highly dynamic network VANET. So in the future paper, we will focus on challenging the actual results in a real world scenario.

8. References [1] H. Wang, R. P. Liu, W. Ni, W. Chen, and I. B. Collings,“Vanet modelling and clustering design under practicaltraffic, channel and mobility conditions,” IEEE Transactionson Communication, 2015. [2] Perkins C.E. and Royer E.M, Ad-hoc, “On-demand Distance Vector Routing”, draft-ietf-manet-aodv02.txt, 1998.

European Journal of Scientific Research - 1450216X, 1450202X http://www.europeanjournalofscientificresearch.com/ [3] https://www.grottonetworking.com/BBD layBlocking.html#queueing-theory

[4] Chao, C.M., Sheu, J.P., and Hu, “Energy conserving gridrouting protocol in mobilead hoc networks,” in IEEE Proceedings of the 2003 International Conference on Parallel Processing (ICPP’03). [5] C. Profentzas, “Studying routing issues in vanets by usingns-3,” Master’s thesis, Alexander Technological EducationalInstitute of Thessaloniki, November 2012. [6] ns-3 Reference Manual / ns-3 project feedback: [email protected] Simulator version: ns-3-dev 20 August 2010. [7] Centre for Applied Informatics (ZAIK), Institute of Transport Research, German Aerospace Centre, Sumo-simulation of urban mobility. http://sumo.sourceforge.net/. [8] BASIC ELEMENTS OF QUEUEING THEORY Application to the Modelling of Computer Systems Philippe NAIN INRIA 2004 route des Lucioles 06902 Sophia Antipolis, France [9] Ueda, F., M. Tanaka, Y. Maeda, M. Namekawa and A. Satoh (2004), Road Traffic Simulation System using the Running Line of Vehicles, Proc. of 23th Simulation Technology Conference, JAPAN, pp.57-60
