Wireless communications using higher frequencies is now possible due to the advancements in the field of high gain antennas. Using such technologies has enabled accessing wireless media within a short range supplying frequency bands with capacity worth multi-gigabits. Higher frequencies are however exposed to blockage events that can hinder the wireless system performance by reducing the throughput and losing user connectivity. The blockage effect becomes more severe with the addition of mobile blockers like vehicles. In order to understand the blockage events induced by a mobile vehicle, an efficient blockage model is required that can assist in the maintenance of the user connection. This paper proposes using a four state channel model based on the user's signal strength for describing the occurrence of blockage events at high frequencies. Signal strength prediction and the channel state classification are then conducted and evaluated using two deep learning neural network disciplines. The high accuracy of the simulation results observed suggest the possibility and implementation of the model in real systems. 

This paper studies the utilization of machine and statistical learning methods for predicting encrypted user traffic. To this end, a reference system model and performance metrics for traffic prediction have been defined for enabling on-line training. Based on a collection of representative traffic data sets including various video and web traffic, two different classes of predictors have been evaluated. Our results show that very good prediction performance can be achieved using long short-term memory (LSTM) recurrent neural networks at the price of a significant computational cost for training. 

Efficient communication and computing resource allocation is becoming a fundamental issue in wireless networks. Efficiency is most often defined in terms of throughput, utilization and spectral efficiency, while the required computational effort is often overlooked. In this paper, we focus on efficient and computationally lightweight downlink scheduling, and we propose a semi-persistent scheduler based on adaptive short term traffic prediction. We evaluate the performance of the proposed scheduler in terms of throughput, fairness, latency, and scheduling complexity. Our numerical results show that scheduling with prediction is a promising approach in improving network performance. The proposed semi-persistent scheduler performs equally well in terms of throughput, fairness, and latency as traditional proportional-fair scheduling, but at a significantly reduced computational cost.

Provision of high data rates with always-onconnectivity to high mobility users is one of the motivations for design of fifth generation (5G) systems. High system capacity can be achieved by coordination between large number of antennas, which is done using the cloud radio access network (CRAN) design in 5G systems. In terms of baseband processing, allocation of appropriate resources to the users is necessary to achieve high system capacity, for which the state of the art uses the users' channel state information (CSI); however, they do not take into account the associated overhead, which poses a major bottleneck for the effective system performance. In contrast to this approach, this paper proposes the use of machine learning for allocating resources to high mobility users using only their position estimates. Specifically, the 'random forest' algorithm, a supervised machine learning technique, is used to design a learning-based resource allocation scheme by exploiting the relationships between the system parameters and the users' position estimates. In this way, the overhead for CSI acquisition is avoided by using the position estimates instead, with better spectrum utilization. While the initial numerical investigations, with minimum number of users in the system, show that the proposed learning-based scheme achieves 86% of the efficiency achieved by the perfect CSI-based scheme, if the effect of overhead is factored in, the proposed scheme performs better than the CSI-based approach. In a realistic scenario, with multiple users in the system, the significant increase in overhead for the CSI-based scheme leads to a performance gain of 100%, or more, by using the proposed scheme, and thus proving the proposed scheme to be more efficient in terms of system performance.

In the context of heterogeneous and small cell networks, users will have the possibility to connect to multiple radio access (RA) carriers that will be available by a dense deployment of radio access infrastructure consisting of high-power and low-power access nodes. Determining which RAs a user should be associated with and select from, for its downlink transmissions, depends on the long-term and short-term data rates that these RAs may offer to the user. In this study the multi-RA allocation and utilisation is decomposed into a multi-RA to user association problem that assigns multiple RAs to users, and a multi-RA selection problem that determines which of the assigned RAs should be used at any time for the user transmissions. As a solution to the first problem, we propose a distributed dual-based spectrum access scheme (DSA) that considers multi-connectivity, whilst, the second problem is solved by means of a heuristic multi-RA selection scheme that utilise different multi-radio transmit diversity (MRTD) schemes while taking into account different inter-cell interference coordination (ICIC) schemes. Our two-step approach is evaluated by means of simulations which demonstrate cell-edge user throughput performance improvements that exceed 100% when the multi-connectivity dual-based spectrum access (MC-DSA) is employed. Further significant user rate and energy efficiency improvements up to 69% and 38% respectively can be achieved when MRTD is combined with ICIC.

Meeting in a sustainable way the demands of an unprecedented explosive growth in the number of connected devices and mobile data traffic, next-generation mobile networks are expected to make flexible use of spectrum and provide higher capacity by a denser deployment of radio access network infrastructure. In addition, the allocation and the utilisation of radio access resources should improve spectrum usage efficiency and keep energy consumption low, implying the capability to effectively exploit the denser deployment of radio accesses and the co-existence of radio accesses belonging to one or multiple radio access technologies. The term radio access (RA) is used to denote the radio resources associated with a frequency carrier that can be allocated to one or multiple users for their data transmissions.

In a network consisting of multiple radio accesses, improvements in spectral and energy efficiency can be achieved when either channel gains are increased or interference is eliminated or both. There are at least three possible approaches to improve performance, namely, (i) by opportunistically utilizing channel conditions of multiple radio accesses, (ii) by mitigating interference across and within radio accesses, and (iii) by redistributing traffic load among radio accesses. In this work, four different technical solutions for downlink transmissions have been studied and evaluated with respect to throughput, spectral efficiency and energy efficiency performance: (a) multi-access transmit diversity, which refers to the dynamic selection of multiple radio accesses for the transmission of a user’s data, (b) inter-cell interference coordination, which mitigates interference by dividing radio access bandwidth among neighbouring nodes, (c) power on/off of access nodes, which mitigates interference by switching off radio accesses of interfering neighbouring nodes, and (d) radio access load balancing, which effectively distributes load by associating users to radio accesses where the expected rate is higher. For the implementation of the technical solutions different algorithms have been devised and their performance have been evaluated for different user distribution scenarios and different heterogeneous multi-radio access networks deployments consisting of at least a tier of macro-cellular and/or a tier of pico-cells of different densities.

The technical solutions are combined into a framework for the allocation and utilisation of radio access resources in heterogeneous multi-radio access dense networks. The framework consists of two subsequent steps: (1) a step solving the multi-radio allocation problem which associates users with a single or multiple radio accesses in the network on the basis of the expected data rates, and (2) a step solving the multi-radio utilisation problem that determines which of the associated radio access(s) should be used at any time for the user data transmissions in the downlink on the basis of the expected instantaneous data rate. To solve the first problem, we employ the flexible spectrum access solution that performs load balancing by associating users to multiple radio accesses while keeping radio accesses without users switched off. For the second problem, we utilise different multi-radio transmit diversity schemes while taking into account different forms of static inter-cell interference coordination. The evaluation of our framework, which is performed by means of simulations, demonstrates significant performance improvements in terms of user throughput, cell-edge throughput, spectral efficiency and energy efficiency.

För att på ett hållbart sätt uppfylla de krav som en explosiv framtida tillväxt i mobildatatrafik medför, förväntas att nästa generations radionät ger en högre nätkapacitet genom en tät utbyggnad av radionätinfrastruktur, och en effektiv användming av radioaccessresurser. Avhandlingen behandlar resursallokering och resursutnyttjande i radionät med flera olika radioaccesser tillhörande en eller flera olika tekniker. Radioaccesserna, som tilldelas radionätets användare kan användas parallellt eller sekvensiellt där bäraren med högst kvalitet används vid varje transmissionstillfälle. Med radioaccess avses här är en frekvensbärare som består av en eller flera radio resurser som kan tilldelas användare för deras dataöverföringar. Fyra olika tekniska lösningar har studerats och utvärderats med avseende på genomströmning, spektrumeffektivitet och energieffektivitet: (a) multi-diversitet, som gäller det dynamiska urvalet av flera radioaccesser för överföring av en användares data, (b) samordning av intercell-interferens, vilket minskar störningar genom att dela upp radiobandbredden bland grannoder, (c) slå på/av transmissionseffekt, vilket minskar störningar genom att stänga av radion i angränsande noder som stör, och (d) lastbalansering mellan radioaccesser, som effektivt fördelar lasten genom att associera användare till radioaccesser där förväntade genomströmningen är högst.

Dessa tekniska lösningar kombineras till ett ramverk för resursallokering och resursutnyttjande i multiradioaccessnät. Ramverket består av en tvåstegslösning till två på varandra följande problem: (i) ett resursallokeringsproblem, där användarna associeras till radioaccesser på grundval av de förväntade genomsnittliga genomströmningarna, och (ii) ett problem för resursutnyttjande, där användarna nyttjar radioaccesserna på grundval av de förväntade momentana genomströmningarna. För att lösa det första problemet, använder vi oss av en teknisk lösning som bygger på lastbalansering, där användare tilldelas en eller flera radioaccesser, samtidigt som radioaccesser utan användare stängs av. Det andra problemet löser vi genom att använda oss av olika former av multi-diversitet, samtidigt som vi tar hänsyn till samordningar av intercellinterferens. För genomförandet av de tekniska lösningarna har olika algoritmer tagits fram och deras prestanda har utvärderats för olika användarscenarier. Vi studerar installationer av heterogena multiradioaccessnät som består av åtminstone en grupp av makroceller och/eller picoceller av olika densitet. Utvärderingen har utförts med simuleringar som visar betydande prestandaförbättringar i form av datagenomströmning, cellkantsgenomströmning, spektrumeffektivitet och energieffektivitet.

The next generation of radio access networks is expected to offer users flexible radio access to a wide range of frequency bands and radio access technologies (RATs) in dense deployments. In this context, we investigate novel lightweight algorithms for spectrum access in ultra-dense networks. We pose the problem as a network utility maximization and apply Lagrange duality to devise distributed algorithms for users to jointly select access nodes and spectrum bands. Leveraging our theoretical framework, we further propose a heuristic scheme for spectrum access based on the expected long-term user data rate per RAT, hence on a measure of the RATs’ traffic load. Numerical examples show that our schemes achieve significant throughput gains on top of the gains achievable by network densification.

The inter-cell interference problem is a key challenge in OFDMA-based cellular networks because it significantly decreases the throughput of the users at the cell-edge. Mitigating this problem requires an inter-cell interference coordination (ICIC) scheme that divides the spectrum into a power constant band and a power limited band for the cell-edge users. Overall, static ICIC or frequency reuse schemes are effective when users are uniformly distributed across the network, whilst non-uniform user distributions suggest semi-static ICIC schemes where the power levels in the power limited bands are effectively set based on the interference situation from neighboring cells. In this paper we define a method that uses path-loss statistical measures to describe the non-uniform distribution between adjacent cells. Based on this method, we devise a new distributed algorithm that derives the suppression levels of transmit power a cell may set at a neighboring cell’s cell-edge band based on the path-loss to the cell-edge users of that neighboring cell. Our algorithm is evaluated by means of simulations and compared to reference frequency reuse schemes based on a non-uniform user distribution scenario where users are distributed in clusters. Simulation results show that the power suppression levels derived by our algorithm greatly improve spectral and energy efficiency as compared to 7-reuse and 1-reuse schemes respectively. It is also shown that our algorithm exploits the nonuniformity in the user distribution to improve the user throughput of cell-edge users without any losses in the user throughput of the cell-centre users.

The inter-cell interference problem is a key challenge for the planning of OFDMA-based cellular networks. This paper shows the possible gains that can be obtained by employing a static Inter-Cell Interference Coordination (ICIC) scheme in the downlink. A simplified simulation-based capacity analysis study is performed investigating the correlation between system capacity and the distributions of power limited frequency bands among the cells. The size of the ICIC area and the power level in the power limited frequency bands have been two of the key configuration parameters that have been studied. Overall, static ICIC mechanisms are effective in scenarios where the user distribution spatially across the network is uniform. To exploit the full potential of coordination we have investigated the system capacity gains taking the non-uniform user distribution into account. With appropriate configuration of power level settings and proper bandwidth allocation to the spatially clustered users, it was shown that there is a great potential for further capacity improvements.

With the advent of multi-radio access (MRA), an integration of differentradio access technologies (RATs) into a heterogeneous radio access network(RAN) becomes feasible. Such integration allows a user to be at any instantof time served by one or multiple radio accesses (RAs) concurrently, where anRA constitutes an independent radio resource, such as a carrier or a channel,implemented by a single RAT. It also allows a user´s service demands to bemapped onto the aggregated network resources in a transparent and efficientway. An approach for the realization of such multi-radio integrated environmentsis through a unifying generic link layer (GLL) that provides joint radiolink processing and enables communication between nodes and devices acrossdifferent radio accesses.Based on the requirements on multi-access, an architecture that supportsMRA is suggested and the functions of GLL that aims at integrating andutilising multiple RATs are defined. We explore the potential for performanceimprovements through novel extensions of the transmission diversityparadigm which builds on GLL functions that enable multi-radio access selection.Multi-radio transmission diversity (MRTD) is defined as the dynamicselection of radio access for the transmission of a user´s data and it can bethought of as consisting of a packet scheduler operating across multiple radiointerfaces. Different MRTD schemes may be envisaged through combinationsof access re-selection rate, transmission parallelism and transmission redundancy.The re-selection rate refers to the rate at which radio access selectionis performed. It may range from multiple IP packets to one single MACframe. Switched MRTD corresponds to an access selection scheme where auser transmits via one RA at a time, while parallel MRTD corresponds to ascheme where simultaneous transmissions over multiple RAs are scheduled.Finally, redundancy refers to the transmission of copies of the same data overmultiple RAs to increase the possibility of correct reception.The benefits of MRTD are investigated by simulation studies on two multiradiocase scenarios, based on generic RATs and on specific RATs respectively.In the RAT generic scenario, switched MRTD has been evaluated for networktopologies of collocated and non-collocated RAs consisting of macroand pico-cells. In the case of collocated RAs, spectral efficiency is increasedby exploiting diversity in multi-path fading while in non-collocated RAs, thespectral efficiency increase is due to diversity exploitation in both shadowingand multi-path fading. Simulation results show that switched MRTD is mostadvantageous when the RAs provide comparable throughputs. Furthermore,when combined with multi-radio ARQ, MRTD significantly reduces packetloss and packet transmission delays. This is also shown in the specific radioaccesssimulation scenario where a delay sensitive voice service is studied. Inaddition, switched MRTD provides comparable gains to parallel MRTD interms of average packet transmission delay and packet loss, while using lessradio resources. In all cases, it is concluded that maximum performance isconditioned on the reporting delays of the channel quality indicator (CQI).Reporting delays of CQI that are half the channel coherence time render sucha complex MRTD mechanism less effective.