| Summary: | Machine Type Communication (MTC) applications were designed to monitor and control elements of our surroundings and environment. MTC applications have a different set of requirements compared to the traditional communication devices, with Machine to Machine (M2M) data being mostly short, asynchronous, bursty and sometimes requiring end-to-end delays below 1ms. With the growth of MTC, the new generation of mobile communications has to be able to present different types of services with very different requirements, i.e. the same network has to be capable of "supplying" connection to the user that just wants to download a video or use social media, allowing at the same time MTC that has completely different requirements, without deteriorating both experiences. The challenges associated to the implementation of MTC require disruptive changes at the Physical (PHY) and Medium Access Control (MAC) layers, that lead to a better use of the spectrum available. The orthogonality and synchronization requirements of the PHY layer of current Long Term Evolution Advanced (LTE-A) radio access network (based on glsofdm and Single Carrier Frequency Domain Equalization (SC-FDE)) are obstacles for this new 5th Generation (5G) architecture. Generalized Frequency Division Multiplexing (GFDM) and other modulation techniques were proposed as candidates for the 5G PHY layer, however they also suffer from visible degradation when the transmitter and receiver are not synchronized, leading to a poor performance when collisions occur in an asynchronous MAC layer. This dissertation addresses the requirements of M2M traffic at the MAC layer applying multipacket reception (MPR) techniques to handle the bursty nature of the traffic and synchronization tones and optimized back-off approaches to reduce the delay. It proposes a new MAC protocol and analyses its performance analytically considering an SC-FDE modulation. The models are validated using a system level cross-layer simulator developed in MATLAB, which implements the MAC protocol and applies PHY layer performance models. The results show that the MAC’s latency depends mainly on the number of users and the load of each user, and can be controlled using these two parameters.
|
|---|