The Research Of AODV-RAI Routing Protocol Based On Location Prediction In Multi-rate Ad Hoc Networks

Ad Hoc network is composed by a group of wireless mobile nodes which will be able to quickly establish instant communication system without relying on the existing fixed infrastructure. It is specialized for centerless, self-organized, the dynamic topology and limited resources of the mobile node. Nowadays, network physical layer uses the IEEE802.11standard to support multiple data rates, Military affairs adopts multi-frequent multi-mode radio station as the military mobile Ad Hoc network nodes, so the study of multi-rate Ad Hoc Networks is more and more in-depth. So far, many different algorithms and protocol designs have been proposed on the topic, they are multi-rate MAC protocol, multi-rate topology control algorithm and very few multi-rate route metrics. Now, the metric algorithm mainly includes Hop-Count, ETX, WCETT, MTM and RAI and so on."Route Assessment Index" metric has the form of entropy function, the selected route is ensured to have less hops, no bottleneck, high reliability and high throughput with the entropy’s properties. Furthermore, an effective route discovery strategy is also introduced. Through sending hello packets periodically, RAI metric statistics the two-way successful transmission rate of a link over a peried of time and detects the maximum possible data rate of link corresponding to the relative distance at the current moment. Then intermediate nodes on-demand and distributed computing the RAI value to select a local optimum. At last, the destination node calculates a complete route RAI value which is used to select the optimal path in the candidate route. This paper describes the implementation of the protocol using RAI in NS2. However, the simple staticing and detecting way, as RAI used, can only used reflect the network status in the past. For mobile network, it can’t reflect the real situation and even can’t represent the future situation, either. In order to slove this problem, this paper proposes a location prediction mechanism based on the ARIMA model. This mechanism acts after the route descovery starting and before the ETX metric taking effect. By optimizing the moving model, it can obtain a more accurate value of the parameter to select the optimal path.This paper describes the implementation of location prediction based RAI (that is LP-RAI) as a route metric and replace the Hop-Count metric for AODV in NS2. And then simulate and compare the performance of two protocols under the network density changing, the node mobile speed changing and the network overhead changing scenes. The simulation result shows that the LP-RAI AODV protocol can increase the throughput and reduce the average transmission delay. Significantly improve network performance.