近年來通訊網路的進步與寬頻網路的發展，使得更多新型的網路服務迅速應運而生，對網路服務的品質(QoS)也更加要求。眾多新興的應用服務中，即時性的應用服務(real-time traffic)，像是Voice over IP (VoIP)以及video on demand (VoD)，對於封包的傳遞延遲時間(delay time)以及抖動(jitter)具有嚴格的要求，router在處理即時性封包時，並非僅將封包快速送出即可，而須在一定的延遲時間內送出。本研究在BBQ﹙Budget-Base QoS﹚的架構下，發展一套簡單而有效的方法，在封包所經過的每一個router時，根據其緊急程度、服務等級及網路處理能力賦予合適的profit function，再對封包重新排程以獲得最大的profit，期望減小在接收端的jitter絕對值總和，進而增進網路的效能。在此我們設計了四種profit function配合Greedy演算法在兩種不同的router架構上調整封包傳送順序，並利用網路模擬工具NS2在即時性訊務分為單一服務等級及雙服務等級的情況下評估我們方法的效能，證明我們的方法較既有的方法更可以有效控制jitter，且能對不同的服務等級做適當的差異化處理以提高整體服務滿意度。 / With fast growth of Internet traffic and applications, real-time streaming applications have attracted significant attention in the current Internet society. Novel real-time Internet applications such as Voice over IP (VoIP) and video on demand (VoD), have stringent delay and jitter requirements. Under these requirements a router not only have to forward real-time packet as fast as possible, it also have to accomplish the job within a time bound. Our research is based on BBQ﹙Budget-Based QoS﹚management architecture. In order to minimize the jitter of real-time services, we propose a new jitter control mechanism that uses per-hop packet scheduling. Four profit function are designed to represent different scheduling preferences. When a packet gets into a router, the router will assign a profit function based on its timeness and the class of service. The scheduler will re-schedule their order in output queue aiming to maximize total profit. Finally we evaluate our proposed mechanism by NS2 simulator in two situations, single real-time service class and two real-time service classes. Simulation results show that by using our approach the performance is better than other approaches. And we investigate the performance of the combination of profit function in handling multiple classes of service.

Wireless Sensor Networks (WSNs) have significant potential in many application domains, and are poised for growth in many markets ranging from agriculture and animal welfare to home and office automation. Although sensor network deployments have only begun to appear, the industry still awaits the maturing of this technology to realize its full benefits. The main constraints to large scale commercial adoption of sensor networks are the lack of available network management and control tools for determining the degree of data aggregation prior to transforming it into useful information, localizing the sensors accurately so that timely emergency actions can be taken at exact location, and scheduling data packets so that data are sent based on their priority and fairness. Moreover, due to the limited communication range of sensors, a large geographical area cannot be covered, which limits sensors application domain. Thus, we investigate a scalable and flexible WSN architecture that relies on multi-modal nodes equipped with IEEE 802.15.4 and IEEE 802.11 in order to use a Wi-Fi overlay as a seamless gateway to the Internet through WiMAX networks. We focus on network management approaches such as sensors localization, data scheduling, routing, and data aggregation for the WSN plane of this large scale multimodal network architecture and find that most existing approaches are not scalable, energy efficient, and fault tolerant. Thus, we introduce an efficient approach for each of localization, data scheduling, routing, and data aggregation; and compare the performance of proposed approaches with existing ones in terms of network energy consumptions, localization error, end-to-end data transmission delay and packet delivery ratio. Simulation results, theoretical and statistical analysis show that each of these approaches outperforms the existing approaches. To the best of our knowledge, no integrated network management solution comprising efficient localization, data scheduling, routing, and data aggregation approaches exists in the literature for a large scale WSN. Hence, we e±ciently integrate all network management components so that it can be used as a single network management solution for a large scale WSN, perform experimentations to evaluate the performance of the proposed framework, and validate the results through statistical analysis. Experimental results show that our proposed framework outperforms existing approaches in terms of localization energy consumptions, localization accuracy, network energy consumptions and end-to-end data transmission delay.

Efficient channel allocation in GSM communication is difficult and believed to be a performance bottleneck. This thesis is dedicated to studying the performance of GSM/EDGE networks in modern usage scenarios (surfing the web, chatting and streaming services) under the assumption that explicit channel allocation is not needed. Users are able to be (re-)allocated at any time without explicit signaling and the typically associated delay. The radio hardware of these users is identical to that of typically used devices and the theoretical peak performance is not affected. This allows the network to use the radio channels in a more efficiently way and the delays from signaling are reduced or eliminated. The thesis studies this problem by statistically modeling such usage scenarios and the resulting channel utilization patterns. The study finds that the latency can be reduced by 40-70% for typical radio hardware, and that network throughput can be improved by up to 20%. However, further studies are needed on the subject.

Over the past few years, the surge in VR (Virtual Reality) video traffic on networks has been remarkable. Nonetheless, a key challenge remains: ensuring a top-notch quality of experience (QoE) for VR video playback, especially when network bandwidth is limited. Prior studies have mainly focused on tile-based adaptive bitrate (ABR) streaming operating at the application layer on the server/client side to improve QoE, using single viewport prediction to conserve bandwidth. However, single-viewpoint prediction models face limitations due to uncertainties linked with head movement, making it difficult to handle sudden user motions effectively. To overcome these constraints, we propose a lightweight multimodal spatial-temporal transformer architecture, which generates multiple viewpoint trajectories and their corresponding probabilities while leveraging historical trajectory information. Consequently, we introduce a multi-agent reinforcement learning (MARL)-based ABR algorithm that capitalizes on multiple viewport prediction for VR video streaming at the application layer. Our algorithm strives to optimize various QoE objectives under diverse network conditions. To address the ABR problem, we formulate it as a Decentralized Partially Observable Markov Decision Process (Dec-POMDP) problem. To tackle this effectively, we develop a MAPPO (Multi-Agent Proximal Policy Optimization) algorithm within a centralized training and decentralized execution (CTDE) framework. Meanwhile, we also improve QoE at the network layer by utilizing network resources in different network nodes during VR video streaming. We present an innovative system called tile-weighted rate-distortion (TWRD) packet scheduling optimization, which takes advantage of viewpoint prediction. The system dynamically assigns weights to tiles and their corresponding packets using the probability of viewpoint prediction. Due to limited bandwidth, the problem of packet scheduling arises, requiring the determination of which packets should be dropped. To address this challenge, we formulate the problem as an optimization task, taking into account error propagation in the video. Our system leverages the weighted rate-distortion information of packets and applies dynamic programming techniques to design an optimal packet scheduling scheme. By selectively dropping packets at network nodes, our proposed system effectively reduces network congestion and enhances the overall performance of VR video streaming systems operating within bandwidth limitations.

隨著智慧家庭概念與技術的興盛與成熟，未來ISP(Internet Service Provider)業者勢必面臨管理大量智慧家庭中各種不同應用競爭頻寬資源的情況。為因應大量且繁雜類型的應用服務彼此競爭智慧家庭端及ISP端的頻寬資源，考量並應用適當的頻寬分配法則以盡可能優化使用者體驗(QoE)是本研究的研究動機。相關文獻的排程演算法如TDPSS (Time Domain Priority Set Scheduler)、MSCDL (Mac Scheduler)、Proportional Fair (PF)及Adaptive Modified Largest Weighted Delay First (AMLWDF)等。若要用以管理大量智慧家庭的頻寬資源時，ISP業者須將家庭申辦頻寬方案以及ISP端的系統頻寬分開考量。ISP在整合(aggregate)多個家庭的頻寬資源請求時，會依服務類別分配頻寬，最後依據不同類別採取適當處理，進而提升不同應用的QoS品質。 本篇論文著重於延遲時間的考量，提出能分類來自各個智慧家庭中，屬於不同 QCI (QoS Class Identifier)級別的頻寬請求並以不同佇列存放，依DADS (Delay Aware Dynamical Scheduling)演算法計算優先權值。我們利用保障頻寬與動態配置頻寬給不同用戶服務佇列，並優先分配頻寬給較高優先權的用戶服務，以期在維持一定公平性的前提下，盡可能地降低延遲來提升QoS品質。 在我們的實驗數據分析中，我們將DADS和其他方法如MSCDL、PF、TDPSS以及AMLWDF進行公平性、產能、延遲以及抖動率等效能優劣的比較與分析。最後在總結與未來研究方向，我們歸納與整理了DADS與PF、MSCDL、TDPSS以及AMLWDF等演算法的效能優劣。實驗結果顯示，在延遲上，DADS勝過PF和TDPSS，但略輸MSCDL及AMLWDF；在抖動率及產能上，DADS均較其它四者為差；公平性上則是劣於TDPSS、PF及MSCDL但優於AMLWDF。雖然DADS在整體的表現並非最好，但在特別重視延遲時間的Category1類別(包含QCI級別為1、2及5的應用服務)的延遲效能僅輸AMLWDF些許，而產能卻明顯勝過AMLWDF，由此可見DADS在Category1的表現最佳。 / With the concept and technology of smart homes becoming more and more mature and popular, Internet service provider (ISP) must face managing large set of various applications from smart homes which competing for bandwidth resources. In order to enhance Quality of Services (QoS) of a lot of various applications while they are competing bandwidth resources of both smart homes (home internal) and Internet service provider (home external), we propose a QoS aware bandwidth allocation criterion to optimize Quality of user Experience (QoE). Since ISP has to manage bandwidth resources of large set of smart homes, in the proposed criterion each ISP separates the bandwidth resources for home external bandwidth and system bandwidth of ISP, respectively. Then, aggregates bandwidth requests of large number of smart homes according to distinct service classes. This thesis focuses on the performance index of delay. We proposed to classify bandwidth requests from smart homes and put them into different queues, finally, calculate priority values by DADS (Delay Aware Dynamical Scheduling) algorithm. The proposed method is able to effectively reduce delay time with certain degree of fairness guarantee by dynamically allocate bandwidth resources for services with distinct service priorities. In the experiments, we compared DADS with other algorithms such as MSCDL (Mac Scheduler), PF (Proportional Fair), TDPSS (Time Domain Priority Set Scheduler) and AMLWDF (Adaptive Modified Largest Weighted Delay First), etc. in terms of fairness, throughput, delay and jitter. The experiments results show that DADS performs much better than PF and TDPSS but a bit worse than MSCDL and AMLWDF in terms of delay. However, DADS shows no better performance than all other algorithms in terms of jitter and throughput. In fairness comparison, DADS is worse than PF, TDPSS and MSCDL but is better than AMLWDF. Though DADS has no superior performance on overall indices, it is a bit worse than AMLWDF in delay of Category1 (including QCI 1, 2 and 5), its throughput is better than AMLWDF. Therefore, DADS’s performance is the best on Category1 considering overall indices.

nÃo hà / O crescimento do nÃmero dos usuÃrios do serviÃo de Voice over IP(VoIP) faz dele o serviÃo com o maior interesse de ser provido por operadoras de telefonia celular. Por outro lado, este demanda um controle de Quality of Service (QoS) bastante rÃgido, o que torna-se mais complicado em redes sem fio, porque alÃm de congestionamentos na rede, os pacotes podem ser perdidos devido à erros nas transmissÃes no enlace de rÃdio. Dentro deste paradigma, estratÃgias de controle de congestionamento aparecem como uma boa soluÃÃo para lidar com as garantias de QoS em situaÃÃes de sobrecarga do sistema, onde os recursos se encontram exauridos e os requerimentos de qualidade se encontram ameaÃados. Este trabalho consiste na avaliaÃÃo de algoritmos de controle de congestionamento objetivando um aumento de capacidade e das garantias de QoS para serviÃos de voz. Os algoritmos avaliados neste trabalho sÃo os escalonamentos de pacotes e os controles de admissÃo. A anÃlise em cenÃrios de serviÃos mistos composto por usuÃrios VoIP e Web tambÃm està contida neste trabalho. O maior foco està no controle do atraso de pacote, jà que este à um requerimento crucial para serviÃos de tempo-real, como o VoIP. Os resultados mostram que um arcabouÃo de controle de congestionamento projetado para este serviÃo à capaz de melhorar o desempenho do sistema e mitigar os efeitos de congestionamento da rede. No cenÃrio de serviÃos mistos, os algoritmos sÃo capazes de efetuar reserva de recursos dependendo da prioridade definida para cada serviÃo, levando a um aumento na qualidade percebida pelo serviÃo mais sensÃvel atravÃs de uma leve degradaÃÃo no serviÃo mais robusto. / The growth in the number of Voice over IP(VoIP) users on the internet makes it the service with the highest interest to be provided by cellular operators. On the other hand, it demands very strict Quality of Service (QoS) control, which becomes even more complicated in wireless networks, because packets can be lost due to radio link transmission erros, as well as networks congestion. Within this paradigm, congestion control strategies appear as a good solution to cope with QoS guarantees under high loads, where the resources are exhausted and the service quality is threatened. This works comprises the evaluation of congestion control algorithms aiming to improve system capacity and QoS guarantees for speech users. The evaluated alagorithms within this work are packet scheduling and admission control. The analysys in mixed services scenarios composed of VoIP and Web users is also provid in this works. The main focus of the framework is to control the packet delay, since it is a crucial requirement for real-time services. The results show thata suitable congestion control framework is able to provid perfomace improvements and mitigation of the the effects from overloaded conditions. In the mixed services scenario, the algorithms are capable to perform resource reservation depending on the priority defined to each service, leanding to an increase in the quality of more sensitive service by degrading the more robust service

Progress in energy harvesting technology and the increasing need for the energy efficient and environmentally friendly applications have called for reconsideration of communication systems. This reconsideration results in new problem formulations regarding the recent developments on energy harvesting systems. Recently, optimal strategies for various types of energy harvesting networks have been developed based on different harvesting models. This thesis reports the results of our research to develop the optimal scheduling structures on an energy harvesting broadcast and fading channels, and to devise online implementable algorithms for a point-to-point communication system. Particularly, structural properties of an optimal offline schedule in, (1) an energy harvesting broadcast channel with one transmitter two receivers, (2) a single user communication system under fading conditions, are investigated. Moreover, an online algorithm is proposed for a single-user energy harvesting communication system considering the physical constraints and necessities regarding implementation. The proposed scheme is implemented through GNU Radio framework on a USRP device.

The aim of the thesis is to investigate the design of efficient wireless networks through practical as well as theoretical considerations. We constructed a wireless sensor network (WSN) testbed with battery operated nodes capable of RF communication. The system is a centralized tree-based WSN to study challenges of target modeling, detection, and localization. The testbed employed magnetic sensors, on which relatively few results have been reported in the literature. A ferrous test target is modeled as magnetic dipole by validating experimentally. The problem of sensor sensitivity variation is addressed by including sensitivity estimates in model validation. After reliably detecting the target, maximum-likelihood and least-squares techniques are applied for localization. Practical considerations of constructing a WSN utilizing magnetic sensors addressed. Maximum-lifetime operation of these networks requires joint consideration of sensing and communication. Energy harvesting is promising to overcome this major challenge for energy-constrained systems. In the second part of the thesis, we considered the minimization of transmission completion time for a given number of bits per user in an energy harvesting multiuser communication system, where the energy harvesting instants are known beforehand. The two-user case with achievable rate region having structural properties satisfied by the AWGN Broadcast Channel is studied. It is shown that the optimal scheduler ends transmission to both users at the same time while deferring a nonnegative amount of energy from each energy harvest for later use. The problem is formulated as an optimization problem and solved by exploiting its special structure.

In an energy harvesting communication system, energy is derived from outside sources and becomes partially available at different points in time. The constraints induced by this property on energy consumption plays an active role in the design of efficient communication systems. This thesis focuses on the optimal design of transmission and networking schemes for energy harvesting wireless communication systems. In particular, an energy harvesting transmitter broadcasting data to two receivers in an AWGN broadcast channel assuming that energy harvests and data arrivals occur at known instants is considered. In this system, optimal packet scheduling that achieves minimum delay is analyzed. An iterative algorithm, DuOpt, that achieves the same structural properties as the optimal schedule is proposed. DuOpt is proved to obtain the optimal solution when weaker user data is ready at the beginning. A dual problem is defined and shown to be strictly convex. Taking advantage of the dual problem, uniqueness of the solution of the main problem is proved. Finally, it is observed that DuOpt is almost two orders of magnitude faster than the SUMT (sequential unconstrained minimization technique) algorithm that solves the same problem.

Les systèmes de 3G offrent de nouveaux services support (bearer services) à plus hauts débits pour les modes de transmission "paquet". Ces services vont coexister avec la voix (ou d'autres services temps réels), des scénarios de trafic mixte, voix et données, doivent être envisagés. La norme UMTS permet effectivement aux utilisateurs d'avoir plus d'un service activé simultanément. Les différentes classes de trafic augmentent la complexité de la gestion des ressources radios. Dans ce contexte, deux types de fonctions sont étudiés: l'allocation de TFCI et l'ordonnancement de paquets. Leur impact sur la qualité de service (QoS) ainsi que sur la capacité du système est évalué. On propose des améliorations de ces mécanismes dans le but d'augmenter la capacité du système et par conséquent d'améliorer la QoS des utilisateurs. Les études se restreignent au sens montant, c'est à dire aux transmissions du mobile (User Equipment ou UE) vers le réseau. Un premier mécanisme pour lequel un effort d'amélioration est fait, est l'adaptation du lien radio par variation du débit instantané transmis. On simule le cas d'une transmission multiservice (voix et données). L'UE doit partager un débit global qui lui est alloué entre les différents services activés. Ces derniers sont véhiculés dans des radio bearers (tuyaux supports). À chaque intervalle élémentaire de transmission (Transmission Time Interval, TTI), l'UE sélectionne un sous-débit pour chaque bearer; ceci se fait par la sélection d'un "format de transport" à appliquer pendant la durée TTI. Cette procédure s'effectue dans la couche MAC (Medium Access Control); le résultat de la sélection est une combinaison de formats de transport (Transport Format Combination, TFC) que la couche physique doit utiliser. La procédure, nommée sélection de TFC, permet d'adapter la transmission des différents services aux conditions variables de la propagation radio: elle détermine notablement la performance de transmission. L'algorithme de sélection de TFC est tracé dans ses grandes lignes dans la norme. Un de ses principes est de favoriser le trafic temps réel au détriment des services de données par paquet. Cependant, le trafic temps réel peut être perturbé par le trafic de données sous certaines conditions, en particulier pour les mobiles éloignés de la station de base (Node B). On propose un algorithme de sélection de TFC qui limite ces perturbations et qui offre une plus large zone de couverture aux services temps réels. En plus, il améliore la QoS du service de données et le débit effectif de l'UE sans augmenter sa puissance de transmission. Un autre type d'études concerne l'ordonnancement de paquets entre les différents utilisateurs ou UEs. C'est une procédure qui est contrôlée par la partie fixe du réseau. Nous l'étudions principalement par simulation en considérant divers mécanismes ou variations. Un premier mécanisme est nommé fast Variable Spreading Factor (fast VSF): les UEs distants changent rapidement leur facteur d'étalement (SF) afin de conserver une puissance de transmission constante, ce qui vise à stabiliser l'interférence inter-cellulaire. Un deuxième mécanisme étudié est un accès paquet décentralisé (decentralized mode) utilisant une information en retour sur le niveau global d'interférence dans la cellule. Un troisième mécanisme nommé "fast scheduling" (ordonnancement rapide) raccourcit le cycle d'ordonnancement. Les résultats ont montré que dans le cas de faible ou moyenne charge dans la cellule, le mode décentralisé réduit le délai par paquet jusqu'à 25 %. L'ordonnancement rapide augmente la capacité du système jusqu' à 10%. En plus, il améliore la QoS perçue par les utilisateurs en terme de débit par utilisateur et délai par paquet transmis.