Tesis sobre el tema "Access coordination"

Le réseau d'accès radio centralisé (C-RAN) peut fortement augmenter la capacité des réseaux mobiles. Cependant, la faisabilité de C-RAN est limitée par le débit considérable engendré sur les liaisons de transport, appelées également fronthaul. L'objectif de cette thèse est d'améliorer les performances de C-RAN tout en considérant les limitations du débit sur le frontaul, l'allocation de ressources et l'ordonnancement des utilisateurs.Nous étudions d'abord les séparations fonctionnelles possibles entre les têtes radios distantes (RRH) et les unités de traitement en bande de base (BBU) sur la liaison montante pour réduire le débit de transmission sur le fronthaul : certaines fonctions de couche basse sont déplacées du BBU vers les RRH. Nous fournissons une analyse quantitative des améliorations de performances ainsi obtenues.Nous nous concentrons ensuite sur la transmission coordonnée Multi-point (CoMP) sur le lien descendant. CoMP peut améliorer l'efficacité spectrale mais nécessite une coordination inter-cellule, ce qui est possible uniquement si une capacité fronthaul élevée est disponible. Nous comparons des stratégies de transmission avec et sans coordination inter-cellule. Les résultats de simulation montrent que CoMP doit être préféré pour les utilisateurs situés en bordure de cellule et lorsque la capacité du fronthaul est élevée. Nous en déduisons une stratégie hybride pour laquelle Les utilisateurs sont divisés en deux sous-ensembles en fonction de la puissance du signal. Les utilisateurs situés dans les zones centrales sont servis par un seul RRH avec une coordination simple et ceux en bordure de cellule sont servis en mode CoMP. Cette stratégie hybride constitue un bon compromis entre les débits offerts aux utilisateurs et les débits sur le fronthaul
Centralized/Cloud Radio Access Network (C-RAN) is a promising mobile network architecture, which can potentially increase the capacity of mobile networks while reducing operators¿ cost and energy consumption. However, the feasibility of C-RAN is limited by the large bit rate requirement in the fronthaul. The objective of this thesis is to improve C-RAN performance while considering fronthaul throughput reduction, fronthaul capacity allocation and users scheduling.We first investigate new functional split architectures between Remote Radio Heads (RRHs) and Baseband Units (BBU) on the uplink to reduce the transmission throughput in fronthaul. Some low layer functions are moved from the BBU to RRHs and a quantitative analysis is provided to illustrate the performance gains. We then focus on Coordinated Multi-point (CoMP) transmissions on the downlink. CoMP can improve spectral efficiency but needs tight coordination between different cells, which is facilitated by C-RAN only if high fronthaul capacity is available. We compare different transmission strategies without and with multi-cell coordination. Simulation results show that CoMP should be preferred for users located in cell edge areas and when fronthaul capacity is high. We propose a hybrid transmission strategy where users are divided into two parts based on statistical Channel State Informations (CSIs). The users located in cell center areas are served by one transmission point with simple coordinated scheduling and those located in cell edge areas are served with CoMP joint transmission. This proposed hybrid transmission strategy offers a good trade-off between users¿ transmission rates and fronthaul capacity cost

This thesis involved a research of mechanisms of MAC layer in the ad-hoc networks environment, the ad-hoc networks in the terminology of the standard are called IBSS Independent Basic Service, these type of networks are very useful in real situation where there are not the possibility of display a infrastructure, when there isn’t a network previous planning.

The connection to a new network is one of the different with the most common type of Wireless Local Area Networks (WLAN) that are the ones with infrastructure. The connection is established without the presence of a central station, instead the stations discover the others with broadcast messages in the coverage area of each station. In the context of standard 802.11 networks the communication between the stations is peer to peer, only with one hop. To continue with initiation process is necessary the synchronization between the different stations of his timers.

The other capital mechanism that is treated is the medium access mechanism, to hold a shared and unreliable medium, all the heavy of this issue goes to the distributed coordination function DCF.

In this moment there is an emergent technology, WIMAX or standard IEEE 802.16, like the standard 802.11 is a wireless communication protocol. Some comparison between the MAC layer mechanisms would be realized between these two standards

We are at that moment of network evolution when we have realised that our telecommunication systems should mimic features of human kind, e.g., the ability to understand the medium and take advantage of its changes. Looking towards the future, the mobile industry envisions the use of fully automatised cells able to self-organise all their parameters and procedures. A fully self-organised network is the one that is able to avoid human involvement and react to the fluctuations of network, traffic and channel through the automatic/autonomous nature of its functioning. Nowadays, the mobile community is far from this fully self-organised kind of network, but they are taken the first steps to achieve this target in the near future. This thesis hopes to contribute to the automatisation of cellular networks, providing models and tools to understand the behaviour of these networks, and algorithms and optimisation approaches to enhance their performance. This work focuses on the next generation of cellular networks, in more detail, in the DownLink (DL) of Orthogonal Frequency Division Multiple Access (OFDMA) based networks. Within this type of cellular system, attention is paid to interference mitigation in self-organising macrocell scenarios and femtocell deployments. Moreover, this thesis investigates the interference issues that arise when these two cell types are jointly deployed, complementing each other in what is currently known as a two-tier network. This thesis also provides new practical approaches to the inter-cell interference problem in both macro cell and femtocell OFDMA systems as well as in two-tier networks by means of the design of a novel framework and the use of mathematical optimisation. Special attention is paid to the formulation of optimisation problems and the development of well-performing solving methods (accurate and fast).

Doutoramento em Engenharia Electrotécnica
Interest on using teams of mobile robots has been growing, due to their potential to cooperate for diverse purposes, such as rescue, de-mining, surveillance or even games such as robotic soccer. These applications require a real-time middleware and wireless communication protocol that can support an efficient and timely fusion of the perception data from different robots as well as the development of coordinated behaviours. Coordinating several autonomous robots towards achieving a common goal is currently a topic of high interest, which can be found in many application domains. Despite these different application domains, the technical problem of building an infrastructure to support the integration of the distributed perception and subsequent coordinated action is similar. This problem becomes tougher with stronger system dynamics, e.g., when the robots move faster or interact with fast objects, leading to tighter real-time constraints. This thesis work addressed computing architectures and wireless communication protocols to support efficient information sharing and coordination strategies taking into account the real-time nature of robot activities. The thesis makes two main claims. Firstly, we claim that despite the use of a wireless communication protocol that includes arbitration mechanisms, the self-organization of the team communications in a dynamic round that also accounts for variable team membership, effectively reduces collisions within the team, independently of its current composition, significantly improving the quality of the communications. We will validate this claim in terms of packet losses and communication latency. We show how such self-organization of the communications can be achieved in an efficient way with the Reconfigurable and Adaptive TDMA protocol. Secondly, we claim that the development of distributed perception, cooperation and coordinated action for teams of mobile robots can be simplified by using a shared memory middleware that replicates in each cooperating robot all necessary remote data, the Real-Time Database (RTDB) middleware. These remote data copies, which are updated in the background by the selforganizing communications protocol, are extended with age information automatically computed by the middleware and are locally accessible through fast primitives. We validate our claim showing a parsimonious use of the communication medium, improved timing information with respect to the shared data and the simplicity of use and effectiveness of the proposed middleware shown in several use cases, reinforced with a reasonable impact in the Middle Size League of RoboCup.
O interesse na utilização de equipas multi-robô tem vindo a crescer, devido ao seu potencial para cooperarem na resolução de vários problemas, tais como salvamento, desminagem, vigilância e até futebol robótico. Estas aplicações requerem uma infraestrutura de comunicação sem fios, em tempo real, suportando a fusão eficiente e atempada dos dados sensoriais de diferentes robôs bem como o desenvolvimento de comportamentos coordenados. A coordenação de vários robôs autónomos com vista a um dado objectivo é actualmente um tópico que suscita grande interesse, e que pode ser encontrado em muitos domínios de aplicação. Apesar das diferenças entre domínios de aplicação, o problema técnico de construir uma infraestrutura para suportar a integração da percepção distribuída e das acções coordenadas é similar. O problema torna-se mais difícil à medida que o dinamismo dos robôs se acentua, por exemplo, no caso de se moverem mais rápido, ou de interagirem com objectos que se movimentam rapidamente, dando origem a restrições de tempo-real mais apertadas. Este trabalho centrou-se no desenvolvimento de arquitecturas computacionais e protocolos de comunicação sem fios para suporte à partilha de informação e à realização de acções coordenadas, levando em consideração as restrições de tempo-real. A tese apresenta duas afirmações principais. Em primeiro lugar, apesar do uso de um protocolo de comunicação sem fios que inclui mecanismos de arbitragem, a auto-organização das comunicações reduz as colisões na equipa, independentemente da sua composição em cada momento. Esta afirmação é validada em termos de perda de pacotes e latência da comunicação. Mostra-se também como a auto-organização das comunicações pode ser atingida através da utilização de um protocolo TDMA reconfigurável e adaptável sem sincronização de relógio. A segunda afirmação propõe a utilização de um sistema de memória partilhada, com replicação nos diferentes robôs, para suportar o desenvolvimento de mecanismos de percepção distribuída, fusão sensorial, cooperação e coordenação numa equipa de robôs. O sistema concreto que foi desenvolvido é designado como Base de Dados de Tempo Real (RTDB). Os dados remotos, que são actualizados de forma transparente pelo sistema de comunicações auto-organizado, são estendidos com a respectiva idade e são disponibilizados localmente a cada robô através de primitivas de acesso eficientes. A RTDB facilita a utilização parcimoniosa da rede e bem como a manutenção de informação temporal rigorosa. A simplicidade da integração da RTDB para diferentes aplicações permitiu a sua efectiva utilização em diferentes projectos, nomeadamente no âmbito do RoboCup.

Dans les réseaux mobiles du future, un déploiement plus dense des points d’accés radio est prévu pour satisfaire la demande accrue de débit, mais les terminaux utilisateurs peuvent être affectés par une interférence inter-cellulaire plus forte. Par chance, la centralisation des traitements de signal en bande de base dans l’achitecture Cloud RAN (C-RAN) offre la possibilité de la coordination et du traitement conjoint de plusieurs cellules. Pour réellement permettre de déployer ces techniques, une étude bout-à-bout du CRAN est nécessaire selon plusieurs aspects, notamment l’architecture fonctionnelle, la stratégie de coordination, l’implémentation du traitement de signal multiutilisateur et les optimisations possibles pour un fonctionnement plus efficace.Dans cette thèse, nous proposons en premier une architecture qui définit le placement des fonctions du traitement en bande de base entre les unités distribuées et le serveur central. Le but de ce design est de permettre la réalisation des fonctions multi-utilisateurs en transmettant avec la moins de débit possible sur les liens de fronthaul reliant les différentes entités. Dans un second temps, nous présentons comment il est possible de coordiner les différentes cellules servies par le C-RAN en utilisant le concept de réseaux définis par logiciels adapté pour les réseaux d’accès radio. Nous avons mis en place un prototype démontrant la faisabilité de la méthode de contrôle proposée. Finalement, nous étudions l’allocation adaptative du débit sur les liens de fronthaul transportant les symboles numériques quantifiés des utilisateurs en besoin de traitement multi-cellulaire sur la voie montante pour exploiter l’interférence entre eux. Nous proposons un modèle d’optimisation qui inclut le coût des transmissions fronthaul pour maximiser ainsi le gain obtenu par l’opérateur du réseau où la communication multiutilisateur a lieu. Nous réalisons l’optimisation pour différents modèles de coût et en utilisants deux types de données: d’abord les estimations de canal supposées parfaites et disponibles en temps réel, puis seulement les statistiques du canal. Nous montrons que la méthode d’optimisation proposée permet d’exploiter plus efficacement les liens de fronthaul dans l’architecture précedemment définie
In future mobile networks denser deployment of radio access points is planned to satisfy demand of higher throughput, but an increased number of mobile users can suffer from inter-cell interference. Fortunately, the centralization of base-band processing offered by Cloud Radio Access Network (C-RAN) architecture enables coordination and joint physical layer processing between cells. To make practical deployment of these techniques possible, we have to study C-RAN in an end-to-end view regarding several aspects: the functional architecture of a deployment, the multi-cell coordination strategy, the implementation of multi-user signal processing and possibilities for optimization to increase operational efficiency.In this thesis, first, we propose an architecture defining the placement of base-band processing functions between the distributed remote units and the central processing unit. The aim of this design is to enable multi-cell processing both on the uplink and the downlink while requiring low data rate between the involved entities. Secondly, we study how low latency coordination can be realized inside the central unit using software defined networking adapted to radio access networks. Our demonstration through a real-time prototype deployment shows the feasibility of the proposed control framework. Finally, we investigate adaptive allocation of fronthaul rate that is used for transferring quantized base-band symbols for users participating in uplink multi-cell reception in order to exploit interference between them. We propose an optimization model that includes the cost of fronthaul tranmissions and aims to maximize the gain of network operators from multi-user transmissions in C-RAN. We solve the optimization problem for different fronthaul pricing models, in a scenario where real-time and accurate channel estimates are available and in another where only channel statistics are exploited. Using our method - fitting in the architecture that we have defined - cost efficiency of fronthaul usage can be significantly improved

Over the past few decades, wireless local area networks (WLANs) have been widely deployed for data communication in indoor environments such as offices, houses, and airports. In order to fairly and efficiently use the unlicensed frequency band that Wi-Fi devices share, the devices follow a set of channel access rules, which is called a wireless medium access control (MAC) protocol. It is known that wireless devices following the 802.11 standard MAC protocol, i.e. the distributed coordination function (DCF), suffer from packet collisions when multiple nodes simultaneously transmit. This significantly degrades the throughput performance. Recently, several studies have reported access techniques to reduce the number of packet collisions and to achieve a collision free WLAN. Although these studies have shown that the number of collisions can be reduced to zero in a simple way, there have been a couple of remaining issues to solve, such as dynamic parameter adjustment and fairness to legacy DCF nodes in terms of channel access opportunity. Recently, In-Band Full Duplex (IBFD) communication has received much attention, because it has significant potential to improve the communication capacity of a radio band. IBFD means that a node can simultaneously transmit one signal and receive another signal in the same band at the same time. In order to maximize the performance of IBFD communication capability and to fairly share access to the wireless medium among distributed devices in WLANs, a number of IBFD MAC protocols have been proposed. However, little attention has been paid to fairness issues between half duplex nodes (i.e. nodes that can either transmit or receive but not both simultaneously in one time-frequency resource block) and IBFD capable nodes in the presence of the hidden node problem.

Despite the benefits of antiretroviral therapy in treating HIV/AIDS, government responses have varied substantially, from provisions guaranteeing nearly universal access to insufficient provisions providing almost no access. This research seeks to specifically examine primary explanations, such as economic capacity, and emerging explanations, such as the role of electoral accountability and the presence of stigma, and the coordination between the epistemic community and political leadership as potential causes for the variance in the government provision. By controlling for state economic capacity, this research furthers the importance of examining other explanations for state response in light of a public health crisis. While electoral accountability and the role of stigma had marginal impacts, the level of scientific coordination and understanding among the states political leadership had perceptible impacts. This research also tests broader aspects of the political economy such as the role of state capacity and subsequent government crisis response.

We propose a simple, passive and deployable approach for fingerprinting traffic on the wired side as a solution for three critical stealth attacks in wireless networks. We focus on extracting traces of the 802.11 medium access control (MAC) protocol from the temporal arrival patterns of incoming traffic streams as seen on the wired side, to identify attacker behavior. Attacks addressed include unauthorized access points, selfish behavior at the MAC layer and MAC layer covert timing channels. We employ the Bayesian binning technique as a means of classifying between delay distributions. The scheme requires no change to the 802.11 nodes or protocol, exhibits minimal computational overhead and offers a single point of discovery. We evaluate our model using experiments and simulations.

Em ambientes de Computação de Alto Desempenho, as aplicações científicas dependem dos Sistemas de Arquivos Paralelos (SAP) para obter desempenho de Entrada/Saída (E/S), especialmente ao lidar com grandes quantidades de dados. No entanto, E/S ainda é um gargalo para um número crescente de aplicações, devido à diferença histórica entre a velocidade de processamento e de acesso aos dados. Para aliviar a concorrência causada por milhares de nós que acessam um número significativamente menor de servidores SAP, normalmente nós intermediários de E/S são adicionados entre os nós de processamento e o sistema de arquivos. Cada nó intermediário encaminha solicitações de vários clientes para o sistema, uma configuração que dá a este componente a oportunidade de executar otimizações como o escalonamento de requisições de E/S. O objetivo desta dissertação é avaliar diferentes algoritmos de escalonamento, na camada de encaminhamento de E/S, cuja finalidade é melhorar o padrão de acesso das aplicações, agregando e reordenando requisições para evitar padrões que são conhecidos por prejudicar o desempenho. Demonstramos que os escalonadores FIFO (First In, First Out), HBRR (Handle-Based Round-Robin), TO (Time Order), SJF (Shortest Job First) e MLF (Multilevel Feedback) são apenas parcialmente eficazes porque o padrão de acesso não é o principal fator que afeta o desempenho na camada de encaminhamento de E/S, especialmente para requisições de leitura Um novo algoritmo de escalonamento chamado TWINS é proposto para coordenar o acesso de nós intermediários de E/S aos servidores de dados do sistema de arquivos paralelo. Nossa abordagem reduz a concorrência nos servidores de dados, um fator previamente demonstrado como reponsável por afetar negativamente o desempenho. O algoritmo proposto é capaz de melhorar o tempo de leitura de arquivos compartilhados em até 28% se comparado a outros algoritmos de escalonamento e em até 50% se comparado a não fazer o encaminhamento de requisições de E/S.
In High Performance Computing (HPC) environments, scientific applications rely on Parallel File Systems (PFS) to obtain Input/Output (I/O) performance especially when handling large amounts of data. However, I/O is still a bottleneck for an increasing number of applications, due to the historical gap between processing and data access speed. To alleviate the concurrency caused by thousands of nodes accessing a significantly smaller number of PFS servers, intermediate I/O nodes are typically employed between processing nodes and the file system. Each intermediate node forwards requests from multiple clients to the parallel file system, a setup which gives this component the opportunity to perform optimizations like I/O scheduling. The objective of this dissertation is to evaluate different scheduling algorithms, at the I/O forwarding layer, that work to improve concurrent access patterns by aggregating and reordering requests to avoid patterns known to harm performance. We demonstrate that the FIFO (First In, First Out), HBRR (Handle- Based Round-Robin), TO (Time Order), SJF (Shortest Job First) and MLF (Multilevel Feedback) schedulers are only partially effective because the access pattern is not the main factor that affects performance in the I/O forwarding layer, especially for read requests. A new scheduling algorithm, TWINS, is proposed to coordinate the access of intermediate I/O nodes to the parallel file system data servers. Our approach decreases concurrency at the data servers, a factor previously proven to negatively affect performance. The proposed algorithm is able to improve read performance from shared files by up to 28% over other scheduling algorithms and by up to 50% over not forwarding I/O requests.

This paper explores the social aspects of a new kind of mobile games where players interact with each other in a shared physical space as well as in-game. As technology spreads throughout layers of culture and everyday life, and gaming becomes increasingly widespread, we see a future in social digital games through the use of smartphones, because of their prevalence and their technical versatility. This poses new challenges for designers. By using the context of music selection in semi-public to public situations and with the help of prototypes, we explore the problems of making selections and connections in large groups as well as delivering feedback. As a result of this project we arrive at elements such as participation, competition, scalability and the importance of social interactions between participants which can be used when designing systems in similar contexts.

The South African law and legal system can and should be a mechanism through which the lives of all resident there are enhanced through the safeguarding and advancement of the fundamental rights guaranteed in its Bill of Rights. This thesis focuses on ‘legal aid’ service delivery (broadly speaking) for the indigent in a particular locality by and through registered law clinics, other legal services providers and via other means in civil rather than criminal matters. In this regard there continue to be very substantial differences between the proper access to civil justice requirements of constitutional South Africa and the actual situation which has existed since the dawn of South Africa’s democratic era which continues unabated. Justice and equality are promised to all in South Africa, yet due to gaps in the ‘net’ of free legal services provided to the indigent, the ability to pay for legal services in civil cases often remains the deciding factor. This study examines the constitutional obligations which, it is argued, apply to the provision of free civil legal services to impoverished people in South Africa. This research considers the law as a vector for necessary positive transformation in the daily lives of those resident in South Africa, which is considered within the country’s woefully unequal socio-economic situation. It builds upon existing research and court authority which show the function of access to justice as an important promoter of the type of society envisaged by the South African Constitution - one where the enjoyment of justice and equality are within the reach of all. However, at present, if one can afford the expensive services of lawyers in civil matters, then access to justice is far more readily attainable. But the opposite is true where someone is denied meaningful access to justice through a lack of legal representation because they cannot afford prohibitively high lawyers’ costs (and disbursements) and no adequate alternatives are provided for by the state or through other means. In these circumstances a vulnerable, unrepresented litigant in a civil case faces a greatly increased likelihood of being denied proper access to a daunting and intricate legal system. There are two main reasons for concentrating on free legal services to the ‘needy’ in civil rather than criminal matters. In the first place, all available statistics show that a huge proportion of legal aid services in South Africa has been and continues to be dispensed in criminal rather than civil cases. Secondly, there has been minimal research or case authority in South Africa on legal aid and other free legal services for impoverished people in civil matters. This thesis examines the state of free civil legal service provision and the need for such assistance within the eThekwini Metropolitan Municipality, one of South Africa’s largest metropoles. This analysis includes an empirical study of the requests for free civil legal services in a particular year by qualifying potential clients in eThekwini and the degree to which free legal service providers are meeting or failing to meet those needs. The study considers the legal service provision in such matters by legal non-governmental organisations, state-supported legal service providers and the work of legal professionals in private practice acting pro bono. The thesis then proposes a model for eThekwini for coordinating (and concurrently improving) civil legal aid services, pro bono legal work and other forms of free legal assistance - like community service by senior law students and law graduates - in response to the particular needs and circumstances facing the indigent there. When referring to the concept of ‘legal aid services’, this research concentrates on legal advice, assistance and representation to indigent clients. However, the promotion of legal rights awareness to such clients is often necessary to open their eyes to the possibility of legal avenues, where appropriate, to improve their situations. Therefore this study also considers - albeit to a lesser degree - this more indirect form of legal assistance through the dissemination of legal knowledge in an accessible form to clients who would qualify for legal aid assistance. The work concludes by briefly postulating the likely appropriateness (and/or limitations) of the aforementioned ‘free civil legal service model for the indigent’ beyond eThekwini.

La thèse étudie des méthodes d'allocation de ressources, distribuées par station de base (BS) dans les réseaux OFDMA multi-cellulaires. L'objectif est de fournir la Qualité de Service (QdS) requise par chaque utilisateur, quelle que soit sa localisation dans la cellule. Les travaux portent d'abord sur la coordination causale de réseaux. Deux BSs forment un lien MIMO virtuel pour les utilisateurs localisés en bordure de cellule. Ces utilisateurs bénéficient d'un gain de diversité et d'une diminution de l'interférence inter-cellulaire. L'efficacité de la méthode d'allocation de ressources associée dépend de l'équité du contrôle de puissance. En conséquence, la coordination de réseaux est utilisée pour les utilisateurs à Débit Contraint (DC), mais pas pour les utilisateurs Best Effort (BE), dans un algorithme permettant de gérer conjointement les deux objectifs de QdS. La thèse étudie ensuite les réseaux totalement distribués. Pour les utilisateurs DC, une méthode d'allocation de ressources incluant une allocation de puissance itérative est déterminée pour résoudre le problème Margin Adaptive. Cette méthode est étendue aux utilisateurs DC en MIMO, lorsque le transmetteur connaît tout l'information de canal, ou uniquement ses caractéristiques statistiques. Pour les utilisateurs BE, enfin, l'objectif est de maximiser la somme des débits pondérés, le poids de chaque utilisateur étant proportionnel à la longueur de sa file d'attente. Une méthode d'allocation de sous-porteuses, déduite d'un graphe d'interférence, et une méthode de contrôle de puissance distribuée sont proposées pour résoudre ce problème d'optimisation.

This report describes my M.A.Sc. thesis research work. The emerging 4th generation (4G) mobile systems and networks (so called 4G HetNets) are designed as multilayered cellular topology with a number of asymmetrically located, asymmetrically powered, self-organizing, and user-operated indoor small cell (e.g., pico/femto cells and WLANs) with a variety of cell architectures that are overlaid by a large cell (macro cell) with some or all interfering wireless links. These designs of 4G HetNets bring new challenges such as increased dynamics of user mobility and data traffic trespassing over the multi-layered cell boundaries. Traditional approaches of radio resource allocation and inter-cell (cochannel) interference management that are mostly centralized and static in the network core and are carried out pre-hand by the operator in 3G and lower cellular technologies, are liable to increased signaling overhead, latencies, complexities, and scalability issues and, thus, are not viable in case of 4G HetNets. In this thesis a comprehensive research study is carried out on improving the radio resource sharing and inter-cell interference management in 4G HetNets. The solution strategy exploits dynamic and adaptive channel allocation approaches such as dynamic and opportunistic spectrum access (DSA, OSA) techniques, through exploiting the spatiotemporal diversities among transmissions in orthogonal frequency division multiple access (OFDMA) based medium access in 4G HetNets. In this regards, a novel framework named as Hybrid Radio Resource Sharing (HRRS) is introduced. HRRS comprises of these two functional modules: Cognitive Radio Resource Sharing (CRRS) and Proactive Link Adaptation (PLA) scheme. A dynamic switching algorithm enables CRRS and PLA modules to adaptively invoke according to whether orthogonal channelization is to be carried out exploiting the interweave channel allocation (ICA) approach or non-orthogonal channelization is to be carried out exploiting the underlay channel allocation (UCA) approach respectively when relevant conditions regarding the traffic demand and radio resource availability are met. Benefits of CRRS scheme are identified through simulative analysis in comparison to the legacy cochannel and dedicated channel deployments of femto cells respectively. The case study and numerical analysis for PLA scheme is carried out to understand the dynamics of threshold interference ranges as function of transmit powers of MBS and FBS, relative ranges of radio entities, and QoS requirement of services with the value realization of PLA scheme.

碩士
國立雲林科技大學
電機工程系碩士班
94
WLANs are expected to become widely deployed in home-networking environment, and the so-called hot spots. It will provide user some kind of service such as voice, streaming video, Web-browsing data, and data downloading services. In order to provide QoS (Quality of Service) for real-time services, the IEEE 802.11 Task Group E has been working on a proposal for QoS enhancements to the IEEE 802.11 MAC, namely the IEEE 802.11e proposed standard, it will provide QoS guarantee for real-time services over WLAN. IEEE 802.11e has proposed many kinds of mechanisms in the MAC layer to reach the QoS requirements of different services. The major part of the 802.11e is Hybrid Coordination Function (HCF), which can be separated into two access phases: one is HCCA（HCF Controlled Channel Access）, the other is EDCA (Enhanced Distributed Channel Access). The HCCA is polling based and the EDCA is contention base. In this paper, our main objective is to find solutions for applying IEEE 802.11e EDCA in a multiservice environment. First, we give a brief introduction to the history of WLAN and describe the technology and knowledge related to the IEEE 802.11 MAC layer. Second, we illustrate the QoS mechanisms supported by the IEEE 802.11e MAC layer. Third, we present the IEEE 802.11e EDCA analytic model, and investigate the problems of applying EDCA in multiservice environments and their solutions, including fairness problem, EDCA parameter setting, and some problems of coexistence between legacy 802.11 MAC and EDCA mechanisms. Fourth, we give some solutions to integrate the polling base Controlled Access Phase (CAP) into the EDCA contention base to enhance the system performance. Finally, we make our conclusions, and give suggestions for future use of EDCA mechanisms in multiservice environment.

碩士
國立臺灣大學
電機工程學研究所
106
Licensed assisted access (LAA) is the new technology of radio access network (RAN) in 3rd generation partnership project (3GPP) release 13. In addition to accessing licensed band, LAA can also access unlicensed band to increase higher throughput through carrier aggregation. In fact, the demand for spectrum will obviously increase in the future because of 5th generation mobile networks (5G) system such as virtual reality (VR) or the internet of things (IoT). Besides, the phenomenon where people stay in-doors about 80% of the day is another reason. Therefore, combining licensed band and unlicensed band is one of the important ways to satisfy the demand in indoor environ-ment. If every mobile network operator (MNO) wants to fast and widely build high per-formance 5G RAN in indoor, the space will be wasted for same purposes and the oper-ator expense will be expensive. Thus, the concept of micro operator (μO) is proposed. μO will negotiate with the manager of area and build RAN. Then, MNOs can provide services to its subscribers through renting RAN from μO. After that, not only MNOs can reduce its capital expenditure (CAPEX) and operator expense (OPEX) but also μO can get profit from this market. In fact, there is some research about coordination of RAN and spectrum among multiple MNOs by coordinator or μO in indoor environment. However, whether μO will deal with MNOs for renting RAN and how μO coordinate unlicensed band among multiple MNOs have not been discussed. How μO coordinate LAA RAN and spectrum including licensed band unlicensed band among multiple MNOs to provide different quality of service (QoS) for MNOs’ subscribers will be discussed. This research will focus on the incentives and conditions of transaction between MNO and μO and how μO effectively uses unlicensed band. The problem is as following: I. What are the incentives and conditions in which one MNO rents LAA from μO? The challenges lie in the analysis of revenue, cost, permission fee, etc. of LAA for MNO and μO according to the number of subscribers and service scenario. II. How does μO coordinate LAA and unlicensed band between two MNOs? The challenges lie in the rules of allocation on unlicensed band to increase efficiency and fairness. This thesis discusses whether MNO will build LAA network itself or rent LAA net-work from μO according to number of MNO’s subscribers and quantity of spectrum which MNO has. Moreover, to build micro operator opportunity analysis (MOOpA), cost of LAA network, fee of different QoS, heterogeneous spectrum allocation, etc. are considered by paper survey. MOOpA includes MNO transaction module, μO transaction module, media access control of LAA simulator (LAA simulator), etc. to solve the ques-tions in this thesis. In the aspect of cost, CAPEX and OPEX of building one small cell whose lifetime is 10-year are considered. In the aspect of MNO’s revenue, subscribers are divided into two types of QoS by 3GPP corresponding to different fee: Guaranteed bit rate (GBR) and Non-GBR. This thesis follows the rules in 3GPP and implement het-erogeneous spectrum allocation module. In this module, licensed band will be firstly allocated to GBR subscribers because GBR subscribers should be guaranteed with strict Qos. On the other hands, unlicensed band will be firstly allocated to Non-GBR sub-scribers although Non-GBR appropriately use both licensed and unlicensed band, as Non-GBR subscribers only should be guaranteed with loose Qos. To guarantee Qos of every subscriber including guaranteed maximum delay and minimum throughput, pro-portional fair scheduler module, defined by Basukala (2007), is used in LAA simulator to prevent subscribers from “starvation”. Finally, this research uses LAA simulator to evaluate MNO’s revenue in the scenarios of different QoS demands with limited spec-trum MNO having according to quantity of spectrum, number of subscribers, Qos types, fee, cost, etc. Furthermore, we discuss whether μO will deal with MNOs and how much renting RAN is. To coordinate unlicensed band between two MNOs, priority-oriented unlicensed band allocation algorithm module (POUBA) is the new design in the LAA simulator to allocate unlicensed band to MNOs. By the feature of that spectrum can be split equally to N resource blocks (RB), RBs are allocated to MNOs one by one according to the value of MNO’s priority. The value of MNO’s priority has two features as followed: (1) To avoid having idle spectrum, the more remained licensed band, the value of that li-censed band minus the part GBR subscribers use is, the less the value of MNO’s priority is. (2) The more RBs MNO has got, the less the value of MNO’s priority is. Compared with method of equalization where every MNO gets equal quantity of RBs, POUBA can satisfy greater number of Non-GBR subscribers in the scenarios where supply can or cannot meet demand. Eventually, Micro operator coordination of licensed assisted access system is im-plemented in this thesis. The system includes POUBA, graphical user interface (GUI) web of allocation management, database of management, interface ready for setting throughput command on unlicensed band in Industrial Technology Research Institute (ITRI) LAA small cell, etc. Proof of concept of for the feasibility of micro operator co-ordination of licensed assisted access in indoor environment is the goal.

In the last decade, 802.11 wireless devices data-rates have increased by three orders of magnitude, while communications experiencing low throughput are still largely present. Such throughput loss is a fundamental problem of wireless networking that is difficult to diagnose and amend. My research addresses two key causes of throughput loss: MAC layer protocol overhead and destructive link interference. First, I design WiFi-Nano reducing the channel access overhead by an order of magnitude leveraging an innovative speculative technique to transmit preambles. This new concept is based on simultaneous preamble transmission and detection via a self-interference cancellation design, and paves the way to the realization of the collision detection paradigm in wireless networks. Next, I propose 802.11ec (Encoded Control), the first 802.11-based protocol that eliminates the overhead of control packets. Instead, 802.11ec coordinates node transmissions via a set of predefined pseudo-noise codewords, resulting in the dramatic increase of throughput and communication robustness. Finally, I design MIDAS, a model-driven network management tool that alleviates low throughput wireless links identifying key corrective actions. MIDAS' key contribution is to reveal the fundamental role of node transmission coordination in characterizing destructive interference. I implement WiFi-Nano, 802.11ec, and MIDAS using a combination of WARP FPGA-based radio boards, custom emulation platforms, and network simulators. The results obtained show that WiFi-Nano increases the network throughput by up to 100%, 802.11ec improves network access fairness by up to 90%, and MIDAS identifies corrective actions with a prediction error as low as 20%.

碩士
國立臺灣科技大學
電子工程系
93
Voice over Internet Protocol (VoIP) in wireless local area network (WLAN) is an emerging Internet application. VoIP services provide lower cost than Public Swhiched Telephone Network (PSTN). This thesis proposes the Decentralized Coordination Packet Reservation Multiple Access (DC-PRMA) scheme to increase the number of voice calls that can be supported in WLAN. DC-PRMA extends the concept of the Packet Reservation Multiple Access (PRMA) in WLANs. In DC-PRMA, wireless channel is divided into superframes. A superframe is the composite of voice phase and data phase. Voice phase provides reservation resources for voice users. Users in data phase access channel by the Distributed Coordinator Function (DCF) mechanism. Consider the potential hidden terminal problem and multi-rate characteristic in WLAN, an adaptive slot reservation is employed in DC-PRMA. The performance of DC-PRMA is investigated by computer simulations in comparison with legacy IEEE 802.11. The results show that the DC-PRMA scheme can indeed provide more voice calls than legacy IEEE 802.11 while the QoS requirements of voice calls are guaranteed.

博士
國立交通大學
資訊科學與工程研究所
104
Wireless data traffic has seen prolific growth in recent years because the use of smart handheld devices and new emerging services are widespread, such as real-time video streaming and multimedia file sharing. To handle huge wireless data traffic, LTE-A (Long Term Evolution-Advanced) has adopted heterogeneous networks (HetNets) architecture, which consists of macro eNB and pico eNB/relay node (RN), to increase the capacity of LTE-A. In LTE-A HetNets, access control for Machine-to-Machine (M2M) communications, energy saving, and inter-cell interference coordination are three important issues, which are to be resolved in this thesis. For the access control for M2M communications issue, we propose two efficient cooperative access class barring with load balancing (CACB-LB) and traffic adaptive radio resource management (TARRM) schemes for M2M communications. The proposed CACB-LB uses the percentage of the number of MTC devices that can only access one enhanced Node B (eNB) between two adjacent eNBs as a criterion to allocate those MTC devices that are located in the overlapped coverage area to each eNB. Note that an eNB is a base station of LTE-A. In this way, the proposed CACB-LB can achieve better load balancing among eNBs than CACB, which is the best available related work. The proposed CACB-LB also uses the ratio of the channel quality indication (CQI) that an MTC device received from an eNB over the number of MTC devices that attach to the eNB as a criterion to adjust the estimated number of MTC devices that may access the eNB. As a result, the proposed CACB-LB can have a better set of barring rates of access class barring than CACB and can reduce random access delay experienced by an MTC device, which is also applicable to user equipment (UE). In addition, the proposed TARRM allocates radio resources for an MTC device based on the random access rate of the MTC device and the amount of data uploaded and downloaded by the MTC device in a homogeneous MTC device network, and the priority of an MTC device in a heterogeneous MTC device network so as to effectively utilize the radio resources. For the energy saving issue, we propose a self-organizing network (SON)-based adaptive energy saving (AES) mechanism for LTE-A self-organizing HetNets. The proposed AES uses two-level multi-threshold load management for each RN under different eNBs (inter-cell level) and for each RN within the same eNB (intra-cell level) so as to reduce the congestion in hot spot eNBs and RNs. In addition, the proposed AES can dynamically switch an RN between active and sleep modes to maximize the number of sleep RNs for adaptive energy saving. It can also dynamically change an RN’s coverage area to reduce energy consumption and to increase radio resource utilization. Besides, the proposed AES adopts a neural network predictor to forecast the loading of each RN to determine whether it is appropriate to switch an RN to sleep mode. For the inter-cell interference coordination issue, we propose SON-based cell size adaption (SCSA) and traffic adaptive enhanced inter-cell interference coordination (TAeICIC) to resolve the interference problem. The proposed SCSA uses dynamic multi-threshold load management to dynamically set the transmission power of each pico eNB by adjusting the pilot power. In addition, the proposed TAeICIC utilizes a scheduling metric, proportional-fair (PF), which is the estimated throughput based on the CQI reported by a UE divided by the estimated long term average throughput achieved by the UE, to dynamically allocate an appropriate number of Almost Blank Subframes (ABSs) in each ABS period in a macro eNB so as to mitigate the interference from the macro eNB to its adjacent pico eNBs.

This thesis studies the impact of market orientation on the business performance of the Sri Lankan apparel manufacturing organizations in the post MFA environment. An attempt is made to enhance the existing market orientation theory by addressing a research gap and highlighting the importance of the relationship perspective in business to business marketing. A four dimensional market orientation construct is developed to overcome the limitations identified in the extant market orientation theories. The study also identifies resource availability as an antecedent of market orientation along with six other antecedents which are commonly discussed in literature. It further investigates whether the market orientation-business performance relationship is moderated by the prevailing business environment. The proposed modifications to the theory are empirically tested within the context of the apparel industry.

Sector Wide Approach (SWAP) can be defined as collaboration between governments, development partners and civil society. It promotes coordination of various donor funded programmes to align with a common sector vision. In the case of South Africa, a water sector SWAP was introduced in 2000. Funds were directed to municipalities that were providing water known as Water Services Authorities (WSAs). This study examines the impact of the SWAP on tangible water delivery issues in Ukhahlamba District Municipality, Eastern Cape. The literature reviewed focuses on rural development in South Africa, and abroad. The quantitative aspect of the research considered various sources including primary data from municipal records of water services provision. Interviews were held with community members, municipal and provincial government officials. Analysis of information from these sources indicates some aspects of water delivery were positive and some negative. Recommendations for improvement and further research are made in the last chapter.
Development Studies
M.A. (Development Studies)

We develop efficient coordination techniques that support inelastic traffic in large-scale distributed dynamic spectrum access DSA networks. By means of any learning algorithm, the proposed techniques enable DSA users to locate and exploit spectrum opportunities effectively, thereby increasing their achieved throughput (or "rewards" to be more general). Basically, learning algorithms allow DSA users to learn by interacting with the environment, and use their acquired knowledge to select the proper actions that maximize their own objectives, thereby "hopefully" maximizing their long-term cumulative received reward/throughput. However, when DSA users' objectives are not carefully coordinated, learning algorithms can lead to poor overall system performance, resulting in lesser per-user average achieved rewards. In this thesis, we derive efficient objective functions that DSA users an aim to maximize, and that by doing so, users' collective behavior also leads to good overall system performance, thus maximizing each user's long-term cumulative received rewards. We show that the proposed techniques are: (i) efficient by enabling users to achieve high rewards, (ii) scalable by performing well in systems with a small as well as a large number of users, (iii) learnable by allowing users to reach up high rewards very quickly, and (iv) distributive by being implementable in a decentralized manner.
Graduation date: 2013

We have investigated the feasibility of using digital tabletop displays during design team meetings to improve access to design information. To achieve this goal, we have applied various design guidelines for tabletop interface design and requirements for digital meeting scenarios in order to implement design ideas (in a prototype named Mozaic) that specifically cater to design team meetings. Mozaic intends to improve information retrieval and browsing activities during meetings thereby improving information access through a tile based layout, pop-up menu control and filtered list. We have evaluated the features of Mozaic through an exploratory study on a tabletop display (in comparison to a commonly used desktop software tool, which was also used on the tabletop) to evaluate various design ideas and ferret out design ideas for future iterations of Mozaic. We present the result of this study as observations and user feedback and discuss them in the light of interface development for design meeting situations.

Αντικείμενο μελέτης της παρούσας εργασίας αποτελεί ο βαθμός επιρροής της κεφαλής κατά την ανάγνωση παρατακτικών συνθέτων από ασθενείς με neglect dyslexia (ND). Ενώ η πλειοψηφία των συνθέτων της ελληνικής, τα επονομαζόμενα ‘υποτακτικά’ (subordinative compounds), υπακούουν στον νόμο της δεξιόστροφης κεφαλής (Williams, 1981) (π.χ. ελαφο-κυνηγός), τα παρατακτικά σύνθετα (coordinative compounds) δεν εμφανίζουν αυτή την κανονικότητα στη θέση της κεφαλής τους (π.χ αλατο-πίπερο). Κάποιοι θεωρητικοί μορφολόγοι θεωρούν ότι κανένα από τα δύο συστατικά δεν αποτελεί τη βάση του συνθέτου, ούτε μορφολογικά ούτε σημασιολογικά (Ralli, 2005: 174), άλλοι υποστηρίζουν ότι τα παρατακτικά εμφανίζουν δύο κεφαλές, τόσο στο α’ όσο και στο β’ συνθετικό (Kageyama, 2009), ενώ δεν λείπουν και αυτοί που επιχειρηματολογούν για την υπεροχή του δεύτερου συστατικού, λόγω του ότι αυτό καθορίζει το γένος, τον αριθμό και την πτώση των συνθέτων (Guevara & Scalise, 2008). Στο πλαίσιο αυτό, θα προσπαθήσουμε να απαντήσουμε στο ερώτημα της θεωρητικής μορφολογίας σχετικά με τη θέση της κεφαλής στα παρατακτικά [ΟΟ] και [ΕΕ] σύνθετα της Νέας Ελληνικής, καθώς και στο ερώτημα της κλινικής γλωσσολογίας σχετικά με το ρόλο της κεφαλής στην ανάκληση και επεξεργασία των συνθέτων από ασθενείς με neglect dyslexia. Ένα ιδιαίτερα σημαντικό χαρακτηριστικό των ασθενών με ND είναι η ικανότητά τους να αναγνωρίζουν και να σέβονται τα όρια των δύο συστατικών κατά την ανάγνωση των συνθέτων (Behrmann et al., 1990). Επιπλέον, οι Semenza et al. (2011) αναφέρουν για τα Ιταλικά ότι τα λάθη στο α’ συνθετικό είναι πολύ πιο συχνά στα σύνθετα των οποίων η κεφαλή συμπίπτει με το β’ συνθετικό, παρά όταν συμπίπτει με το πρώτο. Ο εξισορροπητικός ρόλος της κεφαλής ως προς την εμφάνιση λαθών πάνω στο συστατικό της, αναμένουμε να έχει αντίκτυπο και στη δική μας περίπτωση: τα υποτακτικά σύνθετα με την κεφαλή στα δεξιά θα πρέπει να εμφανίζουν περισσότερα λάθη στο α’ συνθετικό από τα παρατακτικά, εφόσον τα τελευταία έχουν δύο κεφαλές. Στο πείραμα ανάγνωσης που σχεδιάσαμε, συμπεριλάβαμε 31 παρατακτικά σύνθετα, ονοματικά και επιθετικά (π.χ. ψωμοτύρι ‘bread-cheese’ και κοντόχοντρος αντιστοίχως) και 32 υποτακτικά σύνθετα, επίσης ονοματικά και επιθετικά (π.χ. ερημονήσι και χοντροαλεσμένος αντιστοίχως), τα οποία είχαμε ελένξει ως προς όλες τις σημαντικές ψυχογλωσσολογικές και μορφολογικές παραμέτρους ανά γραμματική κατηγορία. Το πείραμα αυτό το διεξάγαμε σε μία ασθενή 79 ετών, η οποία έχει υποστεί ΔΕ αγγειακό εγκεφαλικό επεισόδιο (ΑΕΕ) και με βάση την επίδοσή της στο BELLS test έχει διαγνωσθεί με left-sided neglect dyslexia. Ενώ στα παρατακτικά σύνθετα της ονοματικής κατηγορίας, η ασθενής τείνει να παραλείπει ή να υποκαθιστά το α΄συστατικό των συνθέτων αυτών εξίσου συχνά με αυτό των υποτακτικών, δεν ισχύει το ίδιο και για τα επιθετικά παρατακτικά, κατά την ανάγνωση των οποίων ο ασθενής κάνει στατιστικά σημαντικά λιγότερα λάθη στο α’ συστατικό απ’ ότι στην ανάγνωση των αντίστοιχων υποτακτικών ( x²= 3.970, p < 0.05). Η συμβολή της παρούσας μελέτης μπορεί να συνοψιστεί στα ακόλουθα σημεία: Πρώτον, αποδεικνύεται ότι οι λεξικοί παράγοντες μπορούν να επηρεάσουν την επιλεκτική προσοχή σε μεγάλο βαθμό. Δεύτερον, παρέχονται τεκμήρια ότι η θεωρητική έννοια της κεφαλής έχει επίδραση στην επεξεργασία των συνθέτων, καθώς η εστίαση της προσοχής επηρεάζεται από και επικεντρώνεται στο συστατικό που λειτουργεί ως κεφαλή, έπειτα από την σιωπηρή ανάγνωση ολόκληρης της λέξης. Τρίτον, ως προς την επίδραση της κεφαλής στα παρατακτικά σύνθετα της ΚΝΕ, από την έρευνά μας προκύπτει ότι ένας διαχωρισμός ανάμεσα στα ονοματικά και τα επιθετικά παρατακτικά σύνθετα, με τα πρώτα να συμπεριφέρονται όπως τα υποτακτικά, ενώ τα δεύτερα να εμφανίζουν μια διευκόλυνση στην ανάγνωση ως απόρροια της ύπαρξης δύο κεφαλών. Στην τελευταία αυτή περίπτωση, ίσως η ανάγνωση επηρεάζεται από την δυνατότητα εναλλαγής της σειράς των όρων ορισμένων επιθετικών παρατακτικών συνθέτων (π.χ. μακρόστενος ~ στενόμακρος), αλλά αυτό είναι κάτι που απαιτεί περαιτέρω μελέτη.
Reading coordinative compound words is investigated in neglect dyslexia in order to assess the influence of ‘headedness’. While the majority of Modern Greek compounds, the so called ‘subordinative compounds’, comply with Williams’ (1981) ‘Righthand Head Rule’ (e.g. elafo-kinigos ‘deer-hunter’), coordinative compounds do not demonstrate clear headedness (e.g. alato-pipero ‘salt-pepper’). Some theoretical morphologists argue that none of the two constituents serves as the basis of the formation, neither morphologically nor semantically (Ralli, 2005: 174), others attest that the formations’ head coincides with their second constituent, because of the second’s constituent inflectional suffix which determine the gender, number and case of the compound (Guevara & Scalise, 2008), while others argue that coordinative compounds are double-headed (Kageyama, 2009). In this context, the theoretical morphology’s question regarding the position of the head in coordinative [NN] and [AA] Greek compounds will be addressed, as well as the question about the function of headedness in compounds’ retrieval and processing by patients suffering from neglect dyslexia. When reading compounds in neglect dyslexia, a common finding is that patients seem to respect the boundaries between the first and the second component (Behrmann et al., 1990). Moreover, Semenza et al. (2011) ascertained that left-headed Italian compounds are read better than right-headed compounds, indicating that the appearance of head on the first constituent, counterbalance the deficit at the processing of this first constituent. According to these findings, it seems reasonable to hypothesize that, if coordinative compounds are indeed headless or double-headed, patients with neglect dyslexia will make fewer mistakes in reading their left constituents compared to subordinative compounds, which have a clear head. A female 79-year old patient, who suffered RH damage and was affected by left-sided neglect dyslexia as diagnosed based on BELLS test, had to read 32 subordinative [NN] and [AA] compounds (e.g. domatosalata ‘tomato-salad’ and piknokatikimenos ‘densely populated’ respectively) and 32 coordinative [NN] and [AA] compounds (e.g. psomotiri ‘bread-cheese’ and glikopikros ‘sweet-bitter’ respectively). Compounds and their constituents were matched for frequency, familiarity, imageability, age of acquisition and orthographic neighbors. Patient performed significantly fewer substitution and omission errors on the left constituent of coordinative compounds when reading compound adjectives (x²= 3.970, p < 0.05), but the same amount of errors in subordinative and coordinative compounds when reading compound nouns. The contribution of the study can be summarized in the following points: First, it showed that lexical factors can influence selective attention to a great extent. Second, it showed that the theoretical concept of headedness does have a processing effect, with the head capturing more attention after implicit reading of the whole word. Third, as for headedness in Greek coordinative compounds, the study revealed a dissociation depending on grammatical class (adjectival vs. nominal compounds) with adjectival coordinative compounds behaving as double-headed while nominal coordinative compounds patterning with subordinative ones. A key factor here might be the interchangeable word order that characterizes adjectival coordinative compounds (e.g makrostenos ‘long-narrow’ ~ stenomakros ‘narrow-long’), but this is something that requires further research.