مسیریابی آگاهانه هدفمند و مکانی شبکه های متحرک ادهاک با Laker LAKER: Location Aided Knowledge Extraction Routing for Mobile Ad Hoc Networks

I. INTRODUCTION Mobile ad hoc network (MANET) is an infrastructureless network formed by a set of wireless nodes that are capable of moving around freely. There is no fixed infrastructure such as basestations. Each mobile node acts as an end-system and a router. Two mobile nodes within transmission range of each other can communicate directly via the ad hoc wireless link. A multihop route is needed when the destination is beyond the coverage of the sender. Hence routing is a key component of MANET performance. A number of routing protocols have been proposed for MANETs during the recent years[1], [2]. Most of these routing protocols can be classified into two categories: proactive protocols and reactive(on-demand) protocols. In proactive approaches, each node will maintain routing information to all possible destinations irrespective of its usage. In on-demand approaches, a node performs route discovery and maintenance only when needed. Due to the nodal mobility and fast changing topology, on demand protocols generally outperform purely proactive protocols. On-demand protocols, such as Dynamic Source Routing (DSR)[4] and Ad hoc On demand Distance Vector (AODV) routing[5], often use flooding techniques to search for a new route. Flooding based route discovery works as follows. When a node S has some data to send to node D but has no existing This research was supported in part by the National Science Foundation through the grants CCR-0296070 and ANI-0296034. route to the destination, it will initiate a route discovery process by broadcasting a route-request packet. An intermediate node I, upon receiving the route-request packet for the first time, will rebroadcast the route-request again if it does not know a route to the destination node D. Finally, when the route-request packet reaches a node(which may be the destination node D itself) that has a route to node D, a route-reply packet is sent back to the sender node S. To reduce the flooding overhead, a variety of optimizations have been developed. For example, DSR aggressively utilizes route caching strategy to reduce the number of route-request messages. As the route-reply message propagates back to the requester, all neighboring nodes along the route can listen to the route information in a promiscuous way and store the route information in its cache. Later, when a new route-request message is propagating in the network, an intermediate node that has a cached route to the destination can reply to the requester without relaying the route-request message. So the total routing overhead can be reduced. The disadvantage of caching route is that these cached routes may be obsolete by the time it is used, especially under relatively high mobility. Our idea is that it will be more desirable to cache some longer-lived properties of the network other than the to-be-broken routes. In real mobility patterns, nodal density may not be uniform across the network. Some parts of the network may cluster many nodes, while some other parts may have sparse nodes. We believe it is useful to cache this kind of nodal distribution information and later use it to guide the route discovery process.