An Autonomous Group Mobility Prediction Model for Simulation of Mobile Ad-hoc through Wireless Network

Group mobility models for ad-hoc wireless networks are frequently used with the purpose of learning the pattern of mobile multi-nodes which move together towards a common destination. Though, network nodes in a mobile Ad-hoc network move in some motion patterns - called mobility models – but there is still a lack of study on the new parameters that are satisfactory for Ad-hoc network partition prediction. Normally, network partitioning is considered as the origin for numerous unexpected serious disruptions in network routing and upper layer applications. Consequently, by exploiting the group mobility pattern, we minimized the extent of such disruptions in an effective way for mobile group users. In this paper, we proposed a new portrayal of group mobility derived from the existing group mobility models available in hand (as secondary data). In general, our model consists of two steps; 1) suggesting a discriminative pattern for group users, and 2) developing a group users’ mobility prediction model. Moreover, in our experiments, we used a simulation technique to improve the performance in terms of precision and recall parameters. In order to take full advantage of significance and prediction of the mobility model, GMP (Group Mobility Pattern) and corruption factors were investigated. The results showed that our method is more accurate in making predictions compared with the two existing methods (MC/MT and TM) designated in this study.