With increasing technological advancements more sophisticated mobile devices are being used by end-users. Third generation (3G) mobile communication systems such as Universal Mobile Telecommunication System (UMTS) are not able to satisfy the rising demand for higher throughputs and low latencies. New standards based on Orthogonal Frequency Division Multiplexing (OFDM), such as Long Term Evolution (LTE) and Worldwide Interoperability for Microwave Access (WiMAX), have been proposed and are currently being integrated into existing mobile networks all over the world. LTE specifications are being finalized within the 3rd Generation Partnership Project (3GPP) with the ambitious goals of increased spectral efficiency and end user throughput. Despite the introduction of several high data rate services, voice communication is still an essential part of the overall wireless wide area cellular communication market. In LTE, the core network is purely packet switched, thus voice is transmitted entirely using a Voice over Internet Protocol (VoIP). Like its predecessor standards it is desired that a large number of simultaneous VoIP calls be supported in LTE, while satisfying the desired Quality of Service (QoS) demands. This thesis examines issues related to VoIP capacity for LTE. One of the key challenges is the limited number of schedulable voice packets per sub frame. The main goal of this thesis is to quantify the impact of this limitation. After describing basic LTE concepts, a detailed description of the control channel resource limitations for the scheduling of voice packets is presented. Consequences of these limitations are explained systematically by presenting the problem in a wider context. Simulation results were obtained using the openWNS Simulator, an event driven system level simulation platform developed at the Communication Networks Research Group (ComNets), RWTH Aachen University Germany. Results are presented showing the impact of different scheduling strategies on VoIP capacity. These results illustrate how the limited control channel resources, specifically the Physical Downlink Control Channel (PDCCH) resources, affect the total number of schedulable VoIP user audio media streams. / Med ökande tekniska framsteg mer avancerade mobila enheter som används av slutanv ändarna. Tredje generationens (3G) mobila kommunikationssystem såsom Universal Mobile Telecommunication System (UMTS) inte kan tillgodose den ökande efterfrågan på högre genomströmning och låga latenser. Nya standarder som bygger på Orthogonal Frequency Division Multiplexing (OFDM), såsom Long Term Evolution (LTE) och Worldwide Interoperability för Microwave Access (WiMAX), har föreslagits och håller på att integreras I befintliga mobilnät över hela världen. LTE specifikationer håller på att färdigställas inom 3rd Generation Partnership Project (3GPP) med de ambitiösa målen om ökad spektral effektivitet och slutanvändare genomstr ömning. Trots införandet av flera tjänster av hög datahastighet, är röstkommunikation fortfarande en väsentlig del av den totala Wireless Wide Area cellulär kommunikation marknaden. I LTE är kärnnätet rent paketförmedlande därmed röst överförs helt och hållet med hjälp av en Voice over Internet Protocol (VoIP). Precis som sina föregångare standarder är det önskvärt att ett stort antal samtidiga VoIP samtal få stöd i LTE, samtidigt som det uppfyller önskade Quality of Service (QoS) krav. Denna avhandling undersöker frågor relaterade till VoIP kapacitet för LTE. En av de viktigaste utmaningarna är det begränsade antalet schemaläggningsbart röst paket per sub ram. Det huvudsakliga målet med denna avhandling är att kvantifiera effekterna av denna begränsning. Efter att ha beskrivit de grundläggande LTE begrepp, är en detaljerad beskrivning av de resurser kontroll kanal begränsningar för schemaläggning av röst paket presenteras. Konsekvenser av dessa begränsningar förklaras systematiskt genom att presentera problemet i ett större sammanhang. Simulering resultat erhölls med hjälp av openWNS Simulator, en händelse driven systemnivå simulering som utvecklats vid Communication Networks Research Group (ComNets), RWTH Aachen University Tyskland. Resultat presenteras som visar effekterna av olika schemaläggning strategier för VoIP kapacitet. Dessa resultat illustrerar hur de begränsade kontroll kanalresurser, särskilt fysiskt Downlink (PDCCH) resurser, påverkar det totala antalet schemaläggningsbart VoIP användare ljud mediaströmmar.

Radio spectrum has become increasingly scarce due to the proliferation of new wireless communication services. This problem has been exacerbated by fixed bandwidth licensing policies that often lead to spectral underutilization. Cognitive radio networks (CRN) can address this issue using flexible spectrum management that permits unlicensed (secondary) users to access the licensed spectrum. Supporting real-time quality-of-service (QoS) in CRNs however, is very challenging, due to the random spectrum availability induced by the licensed (primary) user activity. This thesis considers the problem of real-time voice transmission in CRNs with an emphasis on secondary network ``friendliness''. Friendliness is measured by the secondary real-time voice capacity, defined as the number of connections that can be supported, subject to typical QoS constraints. The constant bit rate (CBR) air interface case is first assumed. An offline scheduler that maximizes friendliness is derived using an integer linear program (ILP) that can be solved using a minimum cost flow graph construction. Two online primary scheduling algorithms are then introduced. The first algorithm is based on shaping the primary spectral hole patterns subject to primary QoS constraints. The second applies real-time scheduling to both primary traffic and virtual secondary calls. The online scheduling algorithms are found to perform well compared to the friendliness upper bound. Extensive simulations of the primary friendly schedulers show the achievable secondary voice capacity for a variety of parameters compared to non-friendly primary scheduling. The thesis then considers the variable bit rate (VBR) air interface option for primary transmissions. Offline and online approaches are taken to generate a primary VBR traffic schedule that is friendly to secondary voice calls. The online VBR schedulers are found to perform well compared to the friendliness upper bound. Simulation results are presented that show the effect of the primary traffic load and primary network delay tolerance on the primary network friendliness level towards potential secondary voice traffic. Finally, secondary user friendliness is considered from an infrastructure deployment point of view. A cooperative framework is proposed, which allows the primary traffic to be relayed by helper nodes using decode-and-forward (DF) relaying. This approach decreases the primary traffic channel utilization, which, in turn, increases the capacity available to potential secondary users. A relay selection optimization problem is first formulated that minimizes the primary channel utilization. A greedy algorithm that assigns relay nodes to primary data flows is introduced and found to perform well compared to the optimum bound. Results are presented that show the primary network friendliness for different levels of primary channel utilization. / Dissertation / Doctor of Philosophy (PhD)