An ANN based QoS aware Adaptive Gateway Discovery Approach for MANET and Infrastructured Networks Integration
Prakash Srivastava and Rakesh Kumar
A mobile ad hoc network (MANET) is an autonomous collection of independent nodes cooperating together to form an infrastructure less network spontaneously. For increasing usability of MANET domain which finds application in natural disaster such as earthquake, military battle field etc, it is also desired to be connected with Internet through Internet gateways. Due to dynamic topology of such networks, there are frequent link failures which results in frequent route establishment and gateway discovery procedure. Hence, network performance degrades due to higher delay latency and routing overheads. Therefore, gateway discovery mechanism is integrated with mobility prediction using artificial neural network to provide a robust and efficient MANET-Internet integration strategy. Most of the existing gateway discovery schemes do not take mobility prediction schemes into account while establishing a route from source node to gateway, therefore they suffer from higher delay latency and packet drops while selecting a particular gateway. To avoid this situation, we have utilized the potential of artificial neural network (ANN) to ascertain the mobility prediction of the nodes involved in establishing the path to the Internet gateway. Besides, our scheme also developed a framework for an efficient adaptive gateway discovery mechanism by determining dynamic proactive area and advertisement periodicity according to actual traffic load and network topology. Consequently, our proposed scheme results in enhanced performance and throughput comparable with other state-of-the-art related schemes. Comprehensive performance evaluation of the proposed approach is done through simulation under varying scenarios of mobility and active source nodes as well as through analytical modeling.
Keywords: ANN, MANET, Quality-of-Service (QoS), Mobility prediction.