This paper optimizes the fronthaul bit allocation in massive multi-user multiple-input multiple-output (MU-MIMO) systems operating with limited-capacity fronthaul links. We consider an advanced antenna system (AAS) controlled by a centralized baseband unit (BBU). In the AAS, multiple antenna elements together with their radio units are integrated into a single unit. In this setup, a key challenge is allocating fronthaul bits between uplink channel state information (CSI) quantization and downlink precoding matrix quantization. We formulate the problem of maximizing the sum spectral efficiency (SE) for a given fronthaul capacity. We develop an SE expression for this scenario based on the hardening bound. We compute the expression in closed form for maximum ratio transmission, which reveals the relative impact of the two types of quantization distortion. We then formulate a bit split optimization problem and propose an algorithm that exactly solves it. Numerical results demonstrate how the relative importance of assigning bits to CSI and precoding varies depending on the signal-to-noise ratio.

Limited fronthaul capacity is a practical bottleneck in massive multiple-input multiple-output (MIMO) 5G architectures, where a base station (BS) consists of an advanced antenna system (AAS) connected to a baseband unit (BBU). Conventional downlink designs perform all precoding at the BBU and transmit a high-dimensional precoding matrix over the fronthaul, resulting in significant quantization loss and signaling overhead. This letter proposes a splitting precoding architecture that separates the design between the AAS and BBU. The AAS performs local subspace selection, after which the BBU computes a quantization-aware refinement precoding over the resulting reduced-dimensional effective channel. Numerical results show that the proposed splitting precoding strategy achieves higher sum rate than conventional one-stage precoding.

In the upcoming era of communication systems, there is an anticipated shift towards using lower-grade hardware components to optimize size, cost, and power consumption. This shift is particularly beneficial for multiple-input multiple-output (MIMO) systems and Internet of Things devices, which require numerous components and extended battery lives. However, using lower-grade components introduces impairments, including various non-linear and time-varying distortions affecting communication signals. Traditionally, these impairments have been treated as additional noise due to the lack of a rigorous theory. This thesis explores a new perspective on how the structure of impairments can be exploited to optimize communication performance. To address these challenges, this thesis presents impairments-aware beamforming in various scenarios. 

Initially, we investigate the systems with limited fronthaul capacity. We propose an optimized linear precoding for advanced antenna systems (AAS) operating at a 5G base station (BS) within the constraints of a limited fronthaul capacity, modeled by a quantizer. The proposed novel precoding minimizes the mean-squared error (MSE) at the receiver side using a sphere decoding (SD) approach. 

After analyzing MSE minimization, a new linear precoding design is proposed to maximize the sum rate of the same system in the second part of this thesis. The latter problem is solved by a novel iterative algorithm inspired by the classical weighted minimum mean square error (WMMSE) approach. Additionally, a quantization-aware low-complexity algorithm expectation propagation (EP) is presented for large massive MIMO setups, which is more practical for nowadays systems. Besides, the heuristic quantization-aware precoding method with lower computational complexity is presented, showing that it outperforms the quantization-unaware baseline. This baseline is an optimized infinite-resolution precoding, which is then quantized. This study reveals that it is possible to double the sum rate at high SNR by selecting weights and precoding matrices that are quantization-aware. 

Next, we adopt a splitting precoding architecture tailored to fronthaul-constrained systems for practical deployments. In modern systems, the AAS can perform part of the beamforming locally, for example, through beam-space selection. The remaining lower-dimensional interference-cancelation precoder can then be transmitted over the limited-capacity fronthaul link. Compared to the previous fully centralized setup under the same fronthaul constraint, this approach enables higher quantization resolution for the precoder coefficients. Moreover, since both the uplink pilot signals used for channel estimation and the downlink precoding matrix must be transmitted over the limited-capacity fronthaul link, we design a joint uplink–downlink bit allocation scheme to determine the optimal distribution of fronthaul resources between the two directions.

In the final part of this thesis, we focus on the signaling problem in mobile millimeter-wave (mmWave) communication. The challenge of mmWave systems is the rapid fading variations and extensive pilot signaling. We explore the frequency of updating the combining matrix in a wideband mmWave point-to-point MIMO under user equipment (UE) mobility. The concept of beam coherence time is introduced to quantify the frequency at which the UE must update its downlink receive combining matrix. The study demonstrates that the beam coherence time can be even hundreds of times larger than the channel coherence time of small-scale fading. Simulations validate that the proposed lower bound on this defined concept guarantees no more than 50 \% loss of received signal gain (SG). Based on these results, beam-coherence-aware two-stage digital combining is proposed for the mmWave single-user point-to-point MIMO and multi-user MIMO systems. We also propose time-domain channel estimation.

I den kommande eran av kommunikationssystem förväntas en förskjutning mot att använda hårdvarukomponenter av lägre kvalitet för att optimera storlek, kostnad och strömförbrukning. Denna förskjutning är särskilt fördelaktig för MIMO-system (multiple-input multiple-output) och sakernas internet-enheter, vilka kräver många komponenter och förlängd batteritid. Användning av komponenter av lägre kvalitet introducerar dock försämringar, inklusive olika icke-linjära och tidsvarierande distorsioner som påverkar kommunikationssignalerna. Traditionellt har dessa försämringar behandlats som ytterligare brus på grund av avsaknaden av en rigorös teori. Denna avhandling utforskar ett nytt perspektiv på hur strukturen av försämringar kan utnyttjas för att optimera kommunikationsprestanda. För att hantera dessa utmaningar presenterar denna avhandling försämringsmedveten strålformning i olika scenarier.

Inledningsvis undersöker vi system med begränsad fronthaul-kapacitet. Vi föreslår en optimerad linjär förkodning för avancerade antennsystem (AAS) som arbetar vid en 5G-basstation (BS) inom begränsningarna av en begränsad fronthaul-kapacitet, modellerad av en kvantiserare. Den föreslagna nya förkodningen minimerar medelkvadratfelet (MSE) på mottagarsidan med hjälp av en sfäravkodningsmetod (SD).

Efter att ha analyserat MSE-minimering föreslås en ny linjär förkodningsdesign för att maximera summahastigheten för samma system i den andra delen av denna avhandling. Det senare problemet löses med en ny iterativ algoritm inspirerad av den klassiska vägda minimum medelkvadratfelsmetoden (WMMSE). Dessutom presenteras en kvantiseringsmedveten lågkomplexitetsalgoritmförväntningsutbredning (EP) för stora massiva MIMO-uppsättningar, vilket är mer praktiskt för dagens system. Dessutom presenteras den heuristiska kvantiseringsmedvetna förkodningsmetoden med lägre beräkningskomplexitet, vilket visar att den överträffar den kvantiseringsomedvetna baslinjen. Denna baslinje är en optimerad förkodning med oändlig upplösning, som sedan kvantiseras. Denna studie visar att det är möjligt att fördubbla summahastigheten vid högt signal-brusförhållande (SNR) genom att välja vikter och förkodningsmatriser som är kvantiseringsmedvetna.

Därefter använder vi en delande förkodningsarkitektur skräddarsydd för fronthaul-begränsade system för praktiska implementeringar. I moderna system kan AAS utföra en del av strålformningen lokalt, till exempel genom strålrymdsval. Den återstående lägre dimensionella interferensutsläckningsförkodaren kan sedan sändas över fronthaul-länken med begränsad kapacitet. Jämfört med den tidigare helt centraliserade uppsättningen under samma fronthaul-begränsning möjliggör denna metod högre kvantiseringsupplösning för förkodarkoefficienterna. Eftersom både upplänkspilotsignalerna som används för kanaluppskattning och nedlänksförkodningsmatrisen måste sändas över fronthaul-länken med begränsad kapacitet, utformar vi dessutom ett gemensamt upplänks-nedlänksbitallokeringsschema för att bestämma den optimala fördelningen av fronthaul-resurser mellan de två riktningarna.

I den sista delen av denna avhandling fokuserar vi på signaleringsproblemet i mobil millimetervågskommunikation (mmWave). Utmaningen med mmWave-system är de snabba fadningsvariationerna och den omfattande pilotsignaleringen. Vi utforskar frekvensen för att uppdatera kombinationsmatrisen i en bredbandig mmWave punkt-till-punkt MIMO under användarutrustningsmobilitet (UE). Konceptet strålkoherenstid introduceras för att kvantifiera frekvensen med vilken UE:n måste uppdatera sin nedlänksmottagningskombinationsmatris. Studien visar att strålkoherenstiden kan vara till och med hundratals gånger större än kanalkoherenstiden vid småskalig fädning. Simuleringar bekräftar att den föreslagna nedre gränsen för detta definierade koncept garanterar högst 50 \% förlust av mottagen signalförstärkning (SG). Baserat på dessa resultat föreslås strålkoherensmedveten tvåstegs digital kombinering för mmWave punkt-till-punkt MIMO för enanvändare och fleranvändar-MIMO-system. Vi föreslår även tidsdomänkanalestimering.

This paper considers a millimeter-wave wideband point-to-point MIMO system with fully digital transceivers at the base station and the user equipment (UE), focusing on mobile UE scenarios. A main challenge when building a digital UE combining is the large volume of baseband samples to handle. To mitigate computational and hardware complexity, we propose a novel two-stage digital combining scheme at the UE. The first stage reduces the N_r received signals to N_c streams before baseband processing, leveraging channel geometry for dimension reduction and updating at the beam coherence time, which is longer than the channel coherence time of the small-scale fading. By contrast, the second-stage combining is updated per fading realization. We develop a pilot-based channel estimation framework for this hardware setup based on maximum likelihood estimation in both uplink and downlink. Digital precoding and combining designs are proposed, and a spectral efficiency expression that incorporates imperfect channel knowledge is derived. The numerical results demonstrate that the proposed approach outperforms hybrid beamforming, showcasing the attractiveness of using two-stage fully digital transceivers in future systems.

This paper considers a multi-user multiple-input multiple-output (MU-MIMO) system where the downlink communication between a base station (BS) and multiple user equipments (UEs) is aided by a reconfigurable intelligent surface (RIS). We study the sum rate maximization problem with the objective of finding the optimal precoding vectors and RIS configuration. Due to fronthaul limitation, each entry of the precoding vectors must be picked from a finite set of quantization labels. Furthermore, two scenarios for the RIS are investigated, one with continuous infinite-resolution reflection coefficients and another with discrete finite-resolution reflection coefficients. A novel framework is developed which, in contrast to the common literature that only offers sub-optimal solutions for optimization of discrete variables, is able to find the optimal solution to problems involving discrete constraints. Based on the classical weighted minimum mean square error (WMMSE), we transform the original problem into an equivalent weighted sum mean square error (MSE) minimization problem and solve it iteratively. We compute the optimal precoding vectors via an efficient algorithm inspired by sphere decoding (SD). For optimizing the discrete RIS configuration, two solutions based on the SD algorithm are developed: An optimal SD-based algorithm and a low-complexity heuristic method that can efficiently obtain RIS configuration without much loss in optimality. The effectiveness of the presented algorithms is corroborated via numerical simulations where it is shown that the proposed designs are remarkably superior to the commonly used benchmarks.

Multi-Gbps data rates are achievable in millimeter-wave (mmWave) bands, but a prominent issue is the tiny wavelength that results in rapid fading variations and significant pilot signaling for channel estimation. In this letter, we recognize that the angles of scattering clusters seen from the UE vary slowly compared to the small-scale fading. We characterize the beam coherence time, which quantifies how frequently the UE must update its downlink receive combining matrix. The exact beam coherence time is derived in the single-cluster case, and an achievable lower bound is proposed for the multi-cluster case. These values are determined so that at least half of the received signal gain is maintained in between the combining updates. We demonstrate how the beam coherence time can be hundreds of times larger than the channel coherence time of the small-scale fading.

In the upcoming era of communication systems, there is an anticipated shift towards using lower-grade hardware components to optimize size, cost, and power consumption. This shift is particularly beneficial for multiple-input multiple-output (MIMO) systems and internet-of-things devices, which require numerous components and extended battery lifes. However, using lower-grade components introduces impairments, including various non-linear and time-varying distortions affecting communication signals. Traditionally, these distortions have been treated as additional noise due to the lack of a rigorous theory. This thesis explores new perspective on how distortion structure can be exploited to optimize communication performance. We investigate the problem of distortion-aware beamforming in various scenarios. 

In the first part of this thesis, we focus on systems with limited fronthaul capacity. We propose an optimized linear precoding for advanced antenna systems (AAS) operating at a 5G base station (BS) within the constraints of a limited fronthaul capacity, modeled by a quantizer. The proposed novel precoding minimizes the mean-squared error (MSE) at the receiver side using a sphere decoding (SD) approach. 

After analyzing MSE minimization, a new linear precoding design is proposed to maximize the sum rate of the same system in the second part of this thesis. The latter problem is solved by a novel iterative algorithm inspired by the classical weighted minimum mean square error (WMMSE) approach. Additionally, a heuristic quantization-aware precoding method with lower computational complexity is presented, showing that it outperforms the quantization-unaware baseline. This baseline is an optimized infinite-resolution precoding which is then quantized. This study reveals that it is possible to double the sum rate at high SNR by selecting weights and precoding matrices that are quantization-aware. 

In the third part and final part of this thesis, we focus on the signaling problem in mobile millimeter-wave (mmWave) communication. The challenge of mmWave systems is the rapid fading variations and extensive pilot signaling. We explore the frequency of updating the combining matrix in a wideband mmWave point-to-point MIMO under user equipment (UE) mobility. The concept of beam coherence time is introduced to quantify the frequency at which the UE must update its downlink receive combining matrix. The study demonstrates that the beam coherence time can be even hundreds of times larger than the channel coherence time of small-scale fading. Simulations validate that the proposed lower bound on this defined concept guarantees no more than 50 \% loss of received signal gain (SG).

I den kommande eran av kommunikationssystem finns det en förväntad förändringmot att använda hårdvarukomponenter av lägre kvalitet för att optimera storlek, kostnad och strömförbrukning. Denna förändring är särskilt fördelaktig för MIMO-system(multiple-input multiple-output) och internet-of-things-enheter, som kräver många komponenter och förlängd batteritid. Användning av komponenter av lägre kvalitet medfördock försämringar, inklusive olika icke-linjära och tidsvarierande förvrängningar sompåverkar kommunikationssignaler. Traditionellt har dessa förvrängningar behandlatssom extra brus på grund av avsaknaden av en rigorös teori. Denna avhandling utforskarett nytt perspektiv på hur distorsionsstruktur kan utnyttjas för att optimera kommunikationsprestanda. Vi undersöker problemet med distorsionsmedveten strålformning iolika scenarier.

I den första delen av detta examensarbete fokuserar vi på system med begränsadfronthaulkapacitet. Vi föreslår en optimerad linjär förkodning för avancerade antennsystem (AAS) som arbetar vid en 5G-basstation (BS) inom begränsningarna av en begränsad fronthaulkapacitet, modellerad av en kvantiserare. Den föreslagna nya förkodningen minimerar medelkvadratfelet (MSE) på mottagarsidan med användning av ensfäravkodningsmetod (SD).

Efter att ha analyserat MSE-minimering, föreslås en ny linjär förkodningsdesignför att maximera summahastigheten för samma system i den andra delen av dennaavhandling. Det senare problemet löses av en ny iterativ algoritm inspirerad av denklassiska vägda minsta medelkvadratfel (WMMSE)-metoden. Dessutom presenterasen heuristisk kvantiseringsmedveten förkodningsmetod med lägre beräkningskomplexitet, som visar att den överträffar den kvantiseringsomedvetna baslinjen. Denna baslinje är en optimerad förkodning med oändlig upplösning som sedan kvantiseras. Dennastudie avslöjar att det är möjligt att fördubbla summahastigheten vid hög SNR genomatt välja vikter och förkodningsmatriser som är kvantiseringsmedvetna.

I den tredje delen och sista delen av denna avhandling fokuserar vi på signaleringsproblemet i mobil millimetervågskommunikation (mmWave). Utmaningen medmmWave-system är de snabba blekningsvariationerna och omfattande pilotsignalering.Vi utforskar frekvensen av att uppdatera den kombinerande matrisen i en bredbandsmmWave punkt-till-punkt MIMO under användarutrustning (UE) mobilitet. Konceptet med strålkoherenstid introduceras för att kvantifiera frekvensen vid vilken UE:nmåste uppdatera sin nedlänksmottagningskombinationsmatris. Studien visar att strålkoherenstiden kan vara till och med hundratals gånger större än kanalkoherenstiden försmåskalig fädning. Simuleringar bekräftar att den föreslagna nedre gränsen för dettadefinierade koncept inte garanterar mer än 50 % förlust av mottagen signalförstärkning(SG)

In this paper, we consider a downlink multi-user multiple-input multiple-output (MU-MIMO) communication assisted by a reconfigurable intelligent surface (RIS) and study the precoding and RIS configuration design under practical system constraints. These constraints include the limited-capacity fronthaul at the transmitter side and the finite resolution of RIS elements. We investigate the sum mean squared error (MSE) minimization problem and propose an algorithm based on the block coordinate descent method to optimize the precoding, RIS configuration, and receiver gains. We compute the precoding vectors and RIS configuration using the Schnorr-Euchner sphere decoding (SESD) method which delivers the optimal MSE-minimizing solution. We numerically evaluate the performance of the proposed SESD-based methods and corroborate their effectiveness in improving the system performance.

Reconfigurable intelligent surface (RIS) is a newly-emerged technology that might fundamentally change how wireless networks are operated. Though extensively studied in recent years, the practical limitations of RIS are often neglected when assessing the performance of RIS-assisted communication networks. One of these limitations is that each RIS element is restricted to incur a controllable phase shift to the reflected signal from a predefined discrete set. This paper studies an RIS-assisted multi-user multiple-input multiple-output (MIMO) system, where an RIS with discrete phase shifts assists in simultaneous uplink data transmission from multiple user equipments (UEs) to a base station (BS). We aim to maximize the sum rate by optimizing the receive beamforming vectors and RIS phase shift configuration. To this end, we transform the original sum-rate maximization problem into a minimum mean square error (MMSE) minimization problem and employ the block coordinate descent (BCD) technique for iterative optimization of the variables until convergence. We formulate the discrete RIS phase shift optimization problem as a mixed-integer least squares problem and propose a novel method based on sphere decoding (SD) to solve it. Through numerical evaluation, we show that the proposed discrete phase shift design outperforms the conventional nearest point mapping method, which is prevalently used in previous works.

This paper considers a multi-user multiple-input multiple-output (MU-MIMO) system where the precoding matrix is selected in a baseband unit (BBU) and then sent over a digital fronthaul to the transmitting antenna array. The fronthaul has a limited bit resolution with a known quantization behavior. We formulate a new sum rate maximization problem where the precoding matrix elements must comply with the quantizer. We solve this non-convex mixed-integer problem to local optimality by a novel iterative algorithm inspired by the classical weighted minimum mean square error (WMMSE) approach. The precoding optimization subproblem becomes an integer least-squares problem, which we solve with a new algorithm using a sphere decoding (SD) approach. We show numerically that the proposed precoding technique vastly outperforms the baseline of optimizing an infinite-resolution precoder and then quantizing it. We also develop a heuristic quantization-aware precoding that outperforms the baseline while having comparable complexity.