Font Size: a A A

Routing Algorithm In Ad Hoc Networks Based On Topology Prediction

Posted on:2007-05-06Degree:MasterType:Thesis
Country:ChinaCandidate:H Y GaoFull Text:PDF
GTID:2178360182960631Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Ad hoc networks are used more and more widely in military and civilian applications due to the rapid organizing of networks and the outstanding ability of destruction-resistance and self-recovery. It is the innovative, prior, multi-field crossing and practicability study in wireless Ad hoc networks that makes this research field a very wide exploration space at present, constantly creating countless research chances too.The thesis according to the feature of Ad hoc networks, an algorithm of improved Elman recursive neural networks is proposed that can be used to predict the minimum hops of Ad hoc nodes and thus can predict the movements of the nodes. The simulating results show that this algorithm which is based on the features of Ad hoc networks and improved Elman network is feasible and its learning ability becomes stronger and faster than original Elman algorithm. The predict results can provide useful basis for the routing and the security of the Ad hoc networks.Introducing the protocol of Ad hoc networks and comparing it with others. Ad hoc networks are comprised of wireless nodes, which may be highly mobile. These features lead to frequently rerouting operations and degrading the performance of networks. To address this problem, routing protocols which based on prediction have been designed for operations in Ad hoc networks, and have improved the performance of networks.For the insufficiency of DSDV routing protocol which based on topology, a new algorithm to update and maintain the routing table is proposed based on the predictive model and routing change rule. The new algorithm called routing update algorithm of minimum hops. Based on the routing update algorithm of minimum hops, this thesis proposes a new Ad hoc routing algorithm with mobility prediction called PMHDSDV which is an improvement on DSDV. PMHDSDV predicts topology information at the next time unit using neural networks. Then the proposed algorithm adopts the routing update algorithm of minimum hops to update and maintain the routing table. The performance of a routing protocol is a very important problem in a mobile Ad hoc network. However, as the desire for miniaturization and portability of mobile devices. How to reduce the battery energy consumption has become a new researching hot-point in performance evaluation of Ad hoc routing protocols. The thesis, compare and evaluate the performance of DSDV and PMHDSDV algorithms in terms of the energy consumption, the delay and the delivery ratio. Simulation and related analyses showthat improved DSDV protocol has lower energy consumption, delay and delivery ratio than DSDV.
Keywords/Search Tags:Prediction, Ad Hoc Networks, Routing Protocol, Neural Networks
PDF Full Text Request
Related items