Cluster Based Geo-Routing Protocol Hatem Hamad * hhamad@iugaza.edu.ps Abdessalam Elhabbash * ahabbash@iugaza.edu.ps Yousif Mansour * ymansour@iugaza.edu.ps Fuad Abuowaimer * fowaimer@iugaza.edu.ps * The Islamic University of Gaza/ Computer Engineering Department, Gaza, Palestine Abstract This paper addresses the problem of the overhead resulting from flooding the control packets in mobile ad hoc networks in searching for routes between the source and destination. We propose a location enhanced routing protocol for clustered MANETs based on the cluster based routing protocol (CBRP). Our protocol employs local position information obtained by smart antennas to discover routes and make routing decisions for the clustered MANETs. One of the CHs, named general manager (GM), is assigned the responsibility to maintain the local positions of the other nodes. The GM divides the space into four quarters and periodically sends HELLO messages that reach all the nodes. Then each node sends its location information to the GM when this information is changed. Also, the GM has the responsibility to route data from the source to the destination. Simulation results show enhancing the performance of clustered MANETs by decreasing the control packets overhead. Keywords: MANET, Cluster, smart antenna, CBGRP 1. Introduction A Mobile Ad hoc network (MANET) is a self- organizing multi-hop system of wireless nodes that can communicate with each other without pre-existing infrastructure. This type of networks has been used in several applications such as industrial, commercial, cultural and environmental. Ad hoc networks are characterized by limited battery power, limited bandwidth, frequent network topology changes, and rapid mobility. These characteristics make the design of routing protocols a great challenge [1]. Existing ad hoc routing protocols are classified either as proactive (Table-driven) or reactive protocols (On- demand) [2] [3]. Proactive protocols always know the routing information beforehand through periodic route updates. Each node maintains one or more tables to store routing information and refreshes these tables timely. Each node propagates its tables through the network to maintain a consistent network view. The advantage of proactive protocols is that each node has nearly a complete view about the network. Once a source wishes to transmit data to a destination, it immediately looks up the routing table for the needed route. However, proactive protocols do not perform well in high mobility or large networks since the amount of information maintained in tables becomes large. On the other hand, reactive protocols create routes only when needed by the source. When the source requires sending to the destination, it invokes the route discovery procedure to discover the route. Once the route is found, it is maintained by the route maintenance procedure until the destination becomes inaccessible. Although the source has to wait for node discovery delay, practical experiments show the reactive protocols perform better and more suitable for ad hoc networks than proactive ones. The architecture of the ad hoc networks can be classified into flat or hierarchal architectures [4]. Flat architectures do not define any network structure. They encounter scalability problems especially with the increased network size. In these architectures each node has to maintain information about all nodes in the network which becomes significantly large with increasing the network size [11]. In hierarchal architectures nodes are dynamically grouped into clusters [5] [6] [7]. Each cluster has a representative called cluster head (CH). Every node has to join a cluster. A node that belongs to more than one cluster is called a gateway. Figure 1 shows clustered ad hoc network architecture. Routing traffic between clusters is done by the CH. The CH is responsible on collecting control packets, e.g. route discovery packets, and relaying them. On contrast, in flat architectures each node floods control packets which may overwhelm the network with these packets and consumes the limited capacity. In this paper we present a location-enhanced routing protocol for clustered MANETs that employs smart antennas for estimating nodes’ locations. The rest of this paper is organized as follows. Section II presents a