Mobility prediction for wireless network resource management

User mobility prediction has been studied for various applications under diverse scenarios to improve network performance. Examples include driving habits in cell stations, roaming habits of cell phone users, student movement in campuses, etc. Use of such information enables use to better manage and adapt resources to provide improved Quality Of Service (QoS). In particular, advance reservation of resources at future destinations can provide better service to a mobile user. However, there is a certain cost associated with each of these kinds of prediction schemes. We concentrate on a campus network to predict user movement and demonstrate a better movement prediction which can be used to design the network architecture differently. This user prediction scheme uses a second order Markov chain to capture user movement history to make predictions. This is highly suitable for a campus environment because of its simplicity and can also be extended to several other network architectures.