Summary:
In order to meet the ever-increasing need for bandwidth and to offer ever more services, wherever users are, cellular networks are rapidly evolving towards technologies characterized by an increasingly sophisticated radio interface. For example, while the deployment of 4G networks was just beginning, operators already planned the first updates to LTE-A solutions and 5G technologies are currently receiving active attention.
These rapid changes are motivated by the explosion of mobile traffic, as shown by numerous studies and observations on current networks. Users equipped with smartphones, tablets, and other mobile devices mainly generate this traffic.
However most of models for cellular networks in literature do not take into account mobility of users. Authors who have tried to take into account users' mobility, propose models based on hypotheses like users moving with infinite speed. In this thesis we have developed analytical models for 4G and 5G cellular networks taking into account user mobility in a realistic way. The proposed models were designed to be simple and easy to solve, allowing users and networks performance to be evaluated almost instantaneously.
Our first analysis and results where on the impact of mobility in dense LTE-A networks with small cells. We developed two models to access static users performance in small cell with mixed users (static and mobile users). The first model is based on Markov chains and the second one on Processor-Sharing queue. Our second analysis and results where on LTE/LTE-A macrocells with two coding zones and visited by mobile users. We proposed a model based on queuing theory to study the performance of mobile users in a LTE/LTE-A macrocell with different radio conditions over its coverage area. Then, we have then extended these models to the case of homogeneous cellular networks where cells are statistically identical. These models allowed us to show the positive impact of user mobility on performance in a cell or in a network. Moreover, we showed that this performance gain was not a monotonous function of user mobility, which is an important result showing the impact of hard handover implemented in LTE and LTE-A networks on performance. Finally, we turned our attention to heterogeneous networks with different type of cells and visited by users with different profiles (speed, amount of data to be transferred).