Feature Selection for Mobility Pattern Recognition in Mobile Ad-Hoc Networks

Mobility is one of the most challenging issues in mobile Ad-Hoc networks which has significant impact on performance of variety of network protocols. To deal with this matter the protocol designers should be able to analyze movement behavior of mobile nodes in a particular wireless network. In our previous works we have proposed a simple mobility pattern recognition method which can classify mobility traces into different mobility model classes. The main issue here is finding appropriate features which can classify different mobility traces into mobility classes accurately. In this paper we try to introduce suitable and minimal features to make our mobility pattern recognition method able to distinguish and classify different mobility models more accurate. Simulation results show significant efficiency of our proposed mobility pattern recognition method using appropriate feature sets introduces in this paper.