Section: New Results

Wireless Networks

Participants : Adlen Ksentini, Yassine Hadjadj-Aoul, Bruno Sericola.

Long Term Evolution (LTE) represents the next generation of Cellular networks or 4G. It allows increasing the data rate and hence services that can be proposed to users. A notable part of activity in cellular networks and particularly in LTE, is related to increasing the user QoE. Due to their numerous advantages, current trends show a growing number of femtocell deployments. However, femtocells would become less attractive to the general consumers if they cannot keep up with the service quality that the macro cellular network should provide. Given the fact that the quality of mobile services provided at femtocells depends largely on the level of congestion on the backhaul link, in [71] we introduced a flow mobility/handover admission control method that makes decisions on layer-three handovers from macro network to femtocell network and/or on entire or partial flow mobility between the two networks based on predicted QoS taking into account metrics such as network load/congestion indications and based on predicted QoE metrics. In [70] , we proposed a complete framework that anticipates QoS/QoE (Quality of Experience) degradation and proactively defines policies for LTE-connected cars (UEs) to select the most adequate radio access out of WiFi and LTE. For a particular application, the proposed framework considers the application type, the mobility feature (e.g., speed, user mobility entire/partial path, user final/intermediate destination), and the traffic dynamics over the backhauls of both LTE and WiFi networks in order to predict and allow the UE to select the best network that maximizes user QoE throughout the mobility path.

In [33] ,[23] we considered LTE networks as candidates for hosting the Machine to Machine communication (or Machine Type Communication in the 3GPP jargon). One of the most important problems posed by this kind of traffic is congestion. Congestion concerns all the parts of the network, both the radio and the core networks impacting both the user data and the control planes. In these works, we proposed a congestion aware admission control solution that selectively rejects signaling messages from MTC devices at the radio access network following a probability that is set based on a proportional integrative derivative (PID) controller (from control theory) reflecting the congestion level of a relevant core network node.

Another part of our activities in wireless network are related to energy saving. Indeed, one of the biggest problem today in the wireless world is that wireless devices are battery driven, which reduce their operating lifetime. We addressed the energy issue in wireless network for two different contexts: (i) rich media (such as VoIP) delivery in Wireless LAN; (ii) Wireless Sensor Network (WSN).

In WLAN, mobile stations conserve energy by maximizing the sleep mode periods of the wireless interfaces. Despite of its efficiency, this mode is incompatible with real-time applications and media streaming, like VoIP. In fact, maximizing the sleep mode periods is directly translated into an increased delay, which induces packets losses when exceeding certain thresholds (e.g. buffer overflow and late packet loss), and may severely degrade the perceived user's QoE. We first review a clear state of the art on energy saving for mobiles communication [22] . Then, in [56] , we showed the relation between user QoE and the sleep period in the context of Voice over Wireless Lan (VoWLAN). The system was modeled and controlled using a PID controller, which computes the sleep period enabling to reach a QoE reference value. Thus, we achieved the trade-off between energy consumption and QoE.

On the other hand, Wireless Sensor Networks (WSN) protocols focus primarily on power conservation, because of the limited capacity of the sensor nodes' batteries. In [64] we addressed the case of using radio diversity in WSN (more than one antenna). In this work, we proposed a scheme for radio diversity that can balance, depending on the traffic nature in the network, between minimizing the energy consumption or minimizing the end-to-end delay. The proposed scheme combines the benefit of two metrics, which aim separately to minimize the energy consumption, and to minimize delay when delivering packets to the end-user. In [57] , we worked on the localization problem in WSN by introducing a new way to determine the sensors' residence area. Our new localization algorithm is based on the geometric shape of half-symmetric lens. In [81] we developed a performance analysis of a compression scheme designed to save energy, for specific types of WSN.

In [55] , we presented the DVB-T2 simulation module for OPNET. Note that this module is the only available implementation of DVB-T2 in network simulators.