New Velocity Aware Probabilistic Route Discovery Schemes for Mobile Ad hoc Networks Mustafa Bani Khalaf, Ahmed Y. Al-Dubai and Mourad Abed* School of Computing, Edinburgh Napier University 10 Colinton Road, Edinburgh, EH10 5DT, U.K. E-mail :{ m.banikhalaf, a.al-dubai}@napier.ac.uk} * UVHC, LAMIH, Universite de Valenciennes, France, F-59313 Valenciennes, France E-mail: mourad.abed@univ-valenciennes.fr ABSTRACT: In this paper, we investigate efficient strategies for supporting velocity aware probabilistic route discovery in Mobile Ad hoc Networks (MANETs). MANETs usually use broadcast mechanisms to discover routes between nodes by flooding the network with RREQ packets. Usually, the routes of the high mobility nodes have frequent breakages which result in re-discovering the same routes frequently. Hence, uncontrolled RREQ packets can cause more channel contention and increase packets collision rate. This is well- known as the broadcast storm problem where different probabilistic solutions have been suggested to mitigate its side effect. To the best of our knowledge, this study is the first that considers the velocity vector probabilistic route discover in MANETs. Precisely, this study proposes a new Velocity Aware-Probabilistic (VAP) route discovery model, which can exclude unstable nodes while constructing routes between the source and its destination. Our simulation experiments confirm that our proposed model significantly outperforms existing well known solutions in terms of RREQ packet overhead and link stability. Keywords: Speed, Probability, Counter, AODV, Route Discovery, Broadcast problem. 1. INTRODUCTION MANETs topology changes rapidly and frequently as nodes move freely with no restriction in terms of directions or mobility (i.e. speed and pause time). Data routing and packet dissemination is a challenging task typically for nodes that have high speed and different directions. Wireless links between such nodes experience frequent breakages and expired very often. Hence, re-establishing wireless connection requires flooding the network with a large number of control packets such as Route Replay Packets (RREP) and Route Error Packets (RERR), in addition to the extra RREQ packets. For instance, in Ad hoc On Demand Distance Victor (AODV) protocol [1], route discovery phase swaps the network with RREQ control packets to find the optimal route to the required destination. In some cases, the established route could contain unstable nodes, where link breakage is frequent and affect the overall network performance. In MANETs the broadcast storm during the route discovery phase is a well-known problem [2], which occurs when uncontrolled broadcast mechanism is used to disseminate RREQ packets. Different broadcast mechanisms have been adopted in the literature to mitigate the broadcast problems [2] [3]. One of the most efficient suggested solutions is the probabilistic scheme which demonstrates better performance than other existing solutions. However, most of the suggested probabilistic schemes utilize network density [4], distance [5] or position [6] information, which is available with each node to adjust either the rebroadcast probability, counter threshold or the timer. Unlike other existing probabilistic solutions in this paper, we propose two new probabilistic schemes that utilize both sender and receiver velocity vector to calculate the cosine angle between them, and then to set rebroadcast probability, counter threshold and the timer accordingly. The main aim is to prioritize the transmission of the nodes with similar velocity in order to guarantee that only the most stable nodes participate in the route discovery phase, and to avoid the frequent link breakages phenomena. The rest of the paper is organised as follows. Section 2 introduces related work, problem statement and motivation. Section 3 presents a detailed description of the proposed schemes. Section 4 provides the performance evaluation of our algorithm. Finally, Section 5 concludes this study and outlines our future work. 2. RELATED WORK A probabilistic scheme based on the network density information is suggested to mitigate the broadcast storm in AODV [4]. This scheme divides the nodes into four logical groups of density according to the maximum and minimum network density. The density information is collected by broadcasting HELLO packets every second to construct 1- hop neighbour list at each node. The node then decides in which groups it currently belongs to by comparing its neighbour list with the maximum and the minimum network density threshold AVG threshold , which computed as follows: ∑ = = n i i threshold n N AVG 1 (1) Where n is the number of nodes in the network; Ni is the number of neighbours for node X. Another variation of the density probabilistic scheme is suggested in [7]. In this scheme, rebroadcast probability is set according to the number of duplicated RREQ packets instead of the number of neighbours. Each node counts the number of the same received packet (i.e. c) within a random timer. Upon the timer expiration, the node uses the ratio between the total