This thesis focuses a noncoherent two-user uplink system with each user having a single antenna and a base station equipped with a large number of antennas. It is assumed that small scale channel fading is Rayleigh fading and varies in every one time slot. For such massive MIMO uplink system, we consider an optimal finite-alphabet non-orthogonal multiple access (NOMA) design with each user utilizing nonnegative binary modulation. A fast noncoherent maximum likelihood (ML) detection algorithm for the sum constellation of the two users and a corresponding closed form symbol error probability (SEP) formula are derived. In addition, the lower and upper bounds on SEP are established to quantitatively characterize how quickly SEP decays when the number of base station antennas goes to infinity. Two important concepts: full receiver diversity and geometrical coding gain, are introduced. Particularly for two users and three users systems, with each user transmitting nonnegative binary constellation, we obtain an optimal closed form sum constellation that maximizes both the receiver diversity gain and geometrical coding gain. Computer simulations validateour theoretical analysis and demonstrate that our proposed optimal constellation attains significant performance gains over the currently available constellation design for the same massive MIMO upink system.\ Our future work is to develop an algorithm for devising an optimal AUDCG for the considered system in a more general case. / Thesis / Master of Applied Science (MASc)

In the simple form, a communication system includes a transmitter and a receiver. In the transmitter, it transforms the one-hot vector message to produce a transmitted signal. In general, the transmitter demands restrictions on the transmitted signal. The channel is defined by the conditional probability distribution function. On receiving of the transmitted signal with noise, the receiver appears to apply the transformation to generate the estimate of one hot vector message. We can regard this simplest communication system as a specific case of autoencoder from a deep learning perspective. In our case, autoencoder used to learn the representations of the one-hot vector which are robust to the noise channel and can be recovered at the receiver with the smallest probability of error. Our task is to make some improvements on the autoencoder systems. We propose different schemes depending on the different cases. We propose a method based on optimization of softmax and introduce the L1/2 regularization in MSE loss function for SISO case and MIMO case, separately. The simulation shows that both our optimized softmax function method and L1/2 regularization loss function have a better performance than the original neural network framework. / Thesis / Master of Applied Science (MASc)

Yes / To achieve multiple input multiple output (MIMO) in wireless communication, the orthogonal space-time block coding (OSTBC) is evaluated next. At first, the OSTBC design is extended to include Hadamard matrix, referred to in this work, as traditional Hadamard OSTBC. Next, the Hadamard matrix is imposed on the conventional OSTBC, which is referred to, in this work as, Alamouti-Hadamard STBC. Both the traditional Hadamard OSTBC and the conventional STBC are compared with the Alamouti-Hadamard STBC. It will be shown that imposing the Hadamard conditions over the conventional OSTBC, the performance of the OSTBC 2-transmit antenna scheme can be significantly improved in terms of BER performance. All propositions are well supported with analytical derivations.

Yes / The paper presents a new modeling and design concept of antennas using polar- ization diversity of 2 £ 2 and 3 £ 3 Multiple Input Multiple Outputs (MIMO) system that is proposed for future mobile handsets. The channel capacity is investigated and discussed over Raleigh fading channel and compared to a linear/planner antenna array MIMO channel. The capacity is also discussed over three types of power azimuth spectrums. The results are compared to the constraints capacity limits in which the maximum capacity observed.

In this thesis, we evaluate the performance of multi-user detectors over an uplink using AWGN, Rayleigh flat fading single-input single-output (SISO) and multiple-input multiple-output (MIMO) channel models. First, we review the performance of three multiuser receivers; the decorrelator, the MMSE receiver and the multistage parallel interference cancellation receiver in an AWGN and Rayleigh flat fading SISO channel. Next, the V-BLAST algorithm is reviewed and the error propagation of this algorithm is investigated. Then, the V-BLAST algorithm is combined with multiuser receivers to achieve high channel capacity while sharing the spectral resources over a MIMO channel. A bias reduction technique is considered for multistage parallel interference cancellation receiver on both SISO and MIMO channel. Finally, the effect of channel estimation error and timing delay estimation error is evaluated for MIMO systems with multiple users. / Master of Science

[pt] As comunicações de ondas milimétricas (mmWave) são consideradas uma tecnologia essencial para os sistemas celulares de próxima geração, dado que a enorme largura de banda disponível pode potencialmente fornecer as taxas de vários gigabits por segundo. As técnicas convencionais de pré-codificação e combinação são impraticáveis nos cenários da mmWave devido ao custo de fabricação e ao consumo de energia. As alternativas híbridas foram consideradas uma tecnologia promissora para fornecer um compromisso entre a complexidade do hardware e o desempenho do sistema. Um grande número de projetos de pré-codificadores híbridos têm sido proposto com diferentes abordagens. Uma abordagem possível é procurar minimizar a distância euclidiana entre o pré-decodificador híbrido e o pré-decodificador totalmente digital. No entanto, essa abordagem torna o projeto do pré-codificador híbrido um problema de fatoração da matrices difícil de lidar devido às restrições de hardware dos componentes analógicos. Esta tese de doutorado propõe alguns projetos de pré-codificadores e combinadores híbridos por meio de uma estratégia hierárquica. O problema híbrido de pré-codificação / combinação é dividido em partes analógicas e digitais. Primeiro, o pré-codificador / combinador analógico é projetado. Em seguida, com o pré-codificador / combinador analógico fixo, o précodificador / combinador digital é calculado para melhorar o desempenho do sistema. Além disso, métodos de otimização linear e não linear são empregados para projetar a parte analógica do pré-codificador / combinador. A viabilidade dessas propostas é avaliada usando diferentes técnicas de detecção de dados e analisando o desempenho do sistema em termos de taxa de erros de bits (BER), sum–rate e outras métricas, em cenários internos do mmWave, considerando enlace diretos massivo do MU–MIMO. Além disso, este trabalho propõe um método para encontrar aproximações analíticas bastante restritas ao desempenho obtido no BER. A metodologia proposta exigiria o conhecimento da função densidade de probabilidade (fdp) das variáveis relacionadas que são desconhecidas para os cenários mmWave. Para resolver este problema, são utilizadas as aproximações fdp Gamma. As aproximações analíticas do BER resultaram em diferenças não superiores a 0,5 dB em relação aos resultados da simulação em alto SNR. / [en] Millimeter–wave (mmWave) communications have been regarded as a key technology for the next–generation cellular systems since the huge available bandwidth can potentially provide the rates of multiple gigabits per second. Conventional precoding and combining techniques are impractical at mmWave scenarios due to manufacturing cost and power consumption. Hybrid alternatives have been considered as a promising technology to provide a compromise between hardware complexity and system performance. A large number of hybrid precoder designs have been proposed with different approaches. One possible approach is to search for minimizing the Euclidean distance between hybrid precoder and the full-digital precoder. However, this approach makes the hybrid precoder design becomes a matrix factorization problem difficult to deal due to the hardware constraints of analog components. This doctoral thesis proposes some hybrid precoder and combiners designs through a hierarchical strategy. The hybrid precoding/combining problem is divided into analog and digital parts. First, the analog precoder/combiner is designed. Then, with the analog precoder/combiner fixed, the digital precoder/ combiner is computed to improve the system performance. Furthermore, linear and no-linear optimization methods are employed to design the analog part of the precoder/combiner. The viability of these proposals is evaluated using different data detection techniques and analyzing the system performance in terms of bit error rate (BER), sum rate, and other metrics, in indoor mmWave scenarios considering massive MU-MIMO downlink. Also, this work proposes a method to find fairly tight analytic approximations to the obtained BER performance. The methodology proposed would require the knowledge of the probability density function (pdf) of the variables involved, which are unknown for mmWave scenarios. In order to solve this problem, Gamma pdf approximations are used. The analytic BER approximations resulted in differences no larger than 0.5 dB with respect to the simulation results in high SNR.

The past decade has seen an enormous increase in the number of connected wireless devices, and currently there are billions of devices that are connected and managed by wireless networks. At the same time, the applications that are running on these devices have also developed significantly and became more data rate insatiable. As the number of wireless devices and the demand for a high data rate will always increase, in addition to the growing concern about the energy consumption of wireless communication systems, the future wireless communication systems will have to meet three main requirements. These three requirements are: i) being able to achieve high throughput; ii) serving a large number of users simultaneously; and iii) being energy efficient (less energy consumption). Massive multiple-input multiple-output (MIMO) technology can satisfy the aforementioned requirements; and thus, it is a promising candidate technology for the next generations of wireless communication systems. Massive MIMO technology simply refers to the idea of utilizing a large number of antennas at the base station (BS) to serve a large number of users simultaneously using the same time-frequency resources. The hypothesis behind using a massive number of antennas at the BS is that as the number of antennas increases, the channels become favourable. In other words, the channel vectors between the users and their serving BS become (nearly) pairwisely orthogonal as the number of BS antennas increases. This in turn enables the use of linear processing at the BS to achieve near optimal performance. Moreover, a huge throughput and energy efficiency can be attained due to users multiplexing and array gain. In this thesis, we investigate the performance of massive MIMO systems under different scenarios. Firstly, we investigate the performance of a single-cell multi-user massive MIMO system, in which the channel vectors for the different users are assumed to be correlated. In this aspect, we propose two algorithms for users grouping that aim to improve the system performance. Afterwards, the problem of pilot contamination in multi-cell massive MIMO systems is discussed. Based on this discussion, we propose a pilot allocation algorithm that maximizes the minimum achievable rate in a target cell. Following that, we consider two different scenarios for pilot sequences allocation in multi-cell massive MIMO systems. Lower bounds on the achievable rates are derived for two linear detectors, and the performance under different system settings is analysed and discussed for both scenarios. Finally, two algorithms for pilot sequences allocation are proposed. The first algorithm takes advantage of the multiplicity of pilot sequences over the number of users to improve the achievable rate of edge cell users. While the second algorithm aims to mitigate the negative impact of pilot contamination by utilizing more system resources for the channel estimation process to reduce the inter-cell interference.

Array-to-array, or multiple-input multiple-output (MIMO), links are known to provide extremely high spectral efficiencies in rich multipath environments, such as indoor wireless environments. The selection of a subset of receiver array antennas for a MIMO wireless link has been studied by many as a way to reduce cost and complexity in a MIMO system while providing diversity gain. Combined with a switched multi-beam beamformer, it becomes the beam selection system that can gain high signal-to-interference ratio (SIR) improvement in an interference-imited environment. The objective of this research is to evaluate the performance of low-complexity antenna or beam subset selection methods for small MIMO networks. The types of networks include (1) point-to-point MIMO links with out-of-system interference, (2)multi-user networks with a single, but possibly spatially distributed access point. We evaluate various selection techniques on measured indoor channels, which has not been done before. We propose a new practical selection metric, the peak-to-trough ratio of orthogonal frequency division multiplexing (OFDM) training symbols. We also compare antenna and beam selection on measured indoor channels under more general conditions than has previously been done. Finally, we consider some channel modeling issues associated with beamformers. We investigate the validity of three types of statistical MIMO channel models. A new beamformer is designed based on the ideal of the ``Weichselberger model.'

Storskaliga fleranvändar-MIMO-system, med hundratals basstationsantenner, studeras med allt större intresse både inom akademin och industrin. En anledning är att sådana system kan betjäna flera enantennsanvändare samtidigt över samma tids-frekvens-resurs med fleranvändarförkodning. Det innebär högre datahastigheter och bättre spektral effektivitet. En annan anledning är att basstationens energiförbrukning förväntas avta linjärt med antalet antenner tack vare den ökade antennförstärkningen. För att möjliggöra den stora ökningen av antalet antenner, måste priset per antenn, med dess sändtagarkedja, vara lågt. Vore det möjligt att tillverka basstationsantenner av billiga, massproducerade mobiltelefonskomponenter, som effektförstärkare utan avancerad linearisering, då skulle storskalig fleranvändar-MIMO kunna bli verklighet. Effektförstärkare i mobiltelefoner är generellt anpassade att ha hög verkningsgrad och har, i och med detta, kraftigt olinjära överföringsegenskaper. Det är fördelaktigt att sända signaler med lågt toppvärde genom sådana effektförstärkare, för att undvika svår distortion och för att maximera verkningsgraden genom att endast använda en liten avbackning från arbetspunkten. Konventionellt förkodade signaler har tyvärr högt toppvärde (ca. 10 dB). Detta arbete har undersökt en av Mohammed m.fl. (2013a) föreslagen förkodning för storskalig MIMO som resulterar i sändarsignaler med lågt toppvärde. Det visas att denna förkodning ger signaler med ett toppvärde på 4 dB, och att toppvärdet kan göras godtyckligt litet genom att dessutom begränsa fasvariationen. Ju mer fasen begränsas, desto lägre blir emellertid antennförstärkning. Till exempel om fasvariationen begränsas till π/2, sänks toppvärdet till 2,6 dB, men 2-3 dB högre sändareffekt behövs för att bibehålla samma prestanda eller, likvärdigt, så måste basstationen utrustas med 1,6-2,0 gånger fler antenner. Kontinuerlig fasmodulering som ett sätt att få sändarsignaler med konstant envelopp har studerats kort. Lågtoppvärdesförkodning, där sändarsignalerna ligger innanför en cirkel, föreslås som ett sätt att minska den erfodrade sändareffekten utan att öka toppvärdet märkvärt (<4,5 dB) relativt Mohammeds m.fl. förkodning. Förkodningsalgoritmen som utvecklades för detta fastnade dock i lokala minima, vilket försämrade dess prestanda. Sändareffekten kunde därför endast minskas lite grand (<1 dB) vid höga datahastigheter. En preliminär länkbudget baserad på en enkel effektförstärkarmodell har visat att, med fullständig kanalkännedom och i frekvensplatt fädning, skulle lågtoppvärdesförkodning kunna minska energiförbrukningen med 33 % jämfört med konventionell, linjär förkodning i en basstation med 100 antenner. Analysen antyder att olineariserade klass AB mobiltelefonseffektförstärkare kan vara ett alternativ för storskalig fleranvändar-MIMO-basstationer. / Very-large multi-user MIMO systems, with hundreds of base station antennae, are increasingly attracting attention from both academia and industry. One reason is that such systems can use multi-user precoding to simultaneously serve multiple single-antenna users over the same time-frequency resource. This implies increased data rates and improved spectral efficiency. Another reason is that the energy consumed by the base station is expected to decrease linearly with the number of antennae because of the increasing array gain. To enable the massive increase in the number of antennae, each antenna, together with its tranceiver chain, has to be cheap. If one could manufacture base station antennae using low-cost, mass-produced handset technology, including power amplifiers without advanced linearisation techniques, then very-large multi-user MIMO could become reality. Handset power amplifiers generally aim to be power-efficient, and in doing so often have highly non-linear transfer characteristics. It is of benefit to transmit signals with low peak-to-average ratio (PAR) through such power amplifiers, to avoid excessive distortion and to maximise the power efficiency by only having small operating back-offs. Conventionally precoded signals unfortunately have high PAR (approx. 10 dB). This work has investigated the low-PAR precoding scheme for very-large MIMO proposed by Mohammed et al. (2013a). It is shown that, the transmit signals of this precoding scheme have 4 dB PAR, and that by further limiting the phase variation, the PAR can be made arbitrarily small. However, the more the phase is constrained, the smaller the array gain will be. For example, if the phase variation is limited to π/2, the PAR is lowered to 2.6 dB, but 2-3 dB more transmit power is needed to maintain the same performance or, equivalently, 1.6-2.0 times more antennae are needed at the base station. Continuous phase modulation has briefly been studied as a means of producing constant-envelope transmit signals. Low-PAR precoding, where the transmit signals lie inside a circle, is suggested as a way to decrease the required transmit power without increasing PAR noticeably (<4.5 dB) relative to scheme of Mohammed et al. The algorithm that was developed for this purpose, however got stuck in local minima, which degraded its performance. The transmit power could therefore only be slightly (<1 dB) lowered in the regime of high data rates. A preliminary link budget analysis based on a simplistic model of the power amplifier has shown that, assuming perfect channel state information and frequency-flat fading, low-PAR precoding can reduce energy consumption by 33 % compared to conventional linear precoding in a base station with 100 antennae. The analysis suggests that using unlinearised class AB handset power amplifiers might be a viable option for very-large multi-user MIMO base stations. / 大規模多用戶多輸入多輸出通信系統，即配備上百基站天線的系統，正吸引著學術界及工業界越來越多的關注。其中一個原因是通過多用戶預編碼，該系統可以在同一時頻資源上同時服務多個單天線用戶，有效地增加數據速率及頻譜效率。另一個原因是，跟著陣增益的增加，基站功耗將隨天線數線性遞減。爲了使天線數的極大化可行，每根天線與其收發機的成本必須非常廉價。只有在多天線基站的生產中使用低成本的手機配件，比如不包含複雜線性化技術的功率放大器，大規模多用戶多入多出系統才有可能真正實現。 手機功放通常爲了降低功耗而有著高度非線性傳輸特性。因此，通過這樣的功放更適合傳輸低峯均比的信號以避免過度失真，同時可以在小的運作功率回退下提高功耗效率。傳統預編碼的信號峯均比不巧很高（約10分貝）。本論文研究了由Mohammed等人（2013a）提出的低峯均比預編碼。表明該預編碼的信號有4分貝的峯均比，另外加上相位變化約束信號峯均比可以降到任意小。但相位約束越緊陣增益會隨之減小。譬如約束相位變化小於π/2，峯均比降低到2.6分貝，但需要增加2至3分貝的發射功率保持相同的性能，或增加天線數於1.6至2倍。本文也簡要地描述了恆定包絡信號的連續相位調制，並提出一個預編碼讓傳播信號在一個圓內，以便減少所需要的發射功率，而峯均比也不明顯比Mohammed等人的預編碼大（＜4.5分貝）。爲此設計的算法會陷入局部最優點，從而降低其性能。因此傳輸功率只有在高數據速率場景觀察到稍微減小（＜1分貝）。 一個初步的基於簡單的功放模型的鏈路預算分析表明，假設收發端具有全部的信道狀態信息，並假設頻率平坦衰落，在一台以一百根天線的基站，低峯均比預編碼可以比普通的線性預編碼進一步降低33％功耗。也表示在大規模多用戶多入多出基站中使用非線性的手機功放應該是可行的。 / Mehrbenutzer-MIMO-Systeme mit hunderten von Antennen auf Seite der Basisstation werden mit zunehmendem Interesse sowohl von Universitäten als auch in der Telekommunikationsindustrie erforscht. Ein Vorteil Systeme dieser Art ist die gleichzeitige Versorgung mehreren Benutzer mittels Mehrbenutzervorkodierung über dieselbe Zeit-Frequenz-Ressourcen. Dieses führt zu höheren Datenraten und besserer spektraler Effizienz. Ein weiterer Vorteil, wegen des zunehmenden Antennengewinn, ist der mit der Antennenanzahl erwartete linear abnehmende Energieverbrauch der Basisstation. Um eine große Anzahl von Antennen sinnvoll zu ermöglichen, muss jede individuelle Antenne, mit ihrer zugehörigen Sendeempfängerkette, kostengünstig sein. Wäre es möglich, Antennen einer Basisstation auf kostengünstigen serienmäßig hergestellten Mobiltelefonkomponenten, z.B. Leistungsverstärkern ohne komplexe Linearisierung, aufzubauen, könnten Mehrbenutzer-MIMO-Systeme mit hunderten Antennen wirklich realisiert werden. Mobiltelefonleistungsverstärker sind gewöhnlich eher auf hohen Wirkungsgrade angepasst, deren Übertrangungseigenschaften sind daher stark nichtlinear. Es ist von Vorteil, Signale mit niedrigem Scheitelfaktor durch solche Leistungsverstärker zu übertragen, um übermäßige Verzerrung zu vermeiden und die Wirkungsgrad durch kleineren Backoff vom Arbeitspunkt zu maximieren. Leider haben herkömmlich vorkodierte Signale hohen Scheitelfaktor (ca. 10 dB). Diese Arbeit untersucht die Vorkodierungsmethode von Mohammed u.a. (2013a) zur Verringerung des Scheitelfaktors. Es wird gezeigt, dass die Signale dieser Vorkodierungsmethode haben einen Scheitelfaktor von 4 dB und dass der Scheitelfaktor durch eine zusätzliche Begrenzung der Phasenvariation beliebig klein gemacht werden kann. Je mehr die Phasen begrenzt werden, desto kleiner wird jedoch die Antennenverstärkung. Z.B, wenn die Phasenvariation auf π/2 begrenzt wird, wird der Scheitelfaktor auf 2,6 dB reduziert, aber 2-3 dB höhere Sendeleistung ist benötigt, um die gleiche Datenraten zu behalten, oder, entsprechend, muss die Antennenanzahl um einen Faktor 1,6-2 erhöht werden. Modulation mit stetiger Phase, als eine Methode um Sendesignale mit konstanten Einhüllenden zu bekommen, wird kurz untersucht. Eine Vorkodierungsmethode, wo die Signale innerhalb eines Kreises liegen, wird vorgeschlagen, zur Verringerung der erforderlichen Sendeleistung, ohne den Scheitelfaktor (<4,5 dB), im Vergleich zur Vorkodierung von Mohammed u.a, erkennbar zu erhöhern. Der Algorithmus, der für diesen Zweck entwickelt wurde, fährt aber in lokale Mimima fest, was dessen Leistung vermindern. Die Sendeleistung kann deshalb nur im Bereich hohen Datenraten etwas (<1 dB) gesenkt werden. Eine Kanalgewinnanalyse, die auf einem einfachen Leistungsverstärkermodell beruht, zeigt ansatzweise im Fall perfekter Kanalzustandsinformation und Flachschwunds, dass geeignete Scheitelfaktorvorkodierung in einem Basisstation mit 100 Antennen den Stromverbrauch im Vergleich zu herkömmlicher linearer Vorkodierung um 33 % verringern kann. Die Analyse deutet an, dass die Anwendung unlinearisierter Mobiltelefonleistungsverstärker Klasse AB eine Möglichkeit in Mehrbenutzer-MIMO-Basisstationen mit hunderten von Antennen ist.

In this thesis we have presented an implementation of soft Multi Input Multi Output (MIMO) detection, single tree search algorithm on Graphics Processing Units (GPUs). We have compared its performance on different GPUs and a Central Processing Unit (CPU). We have also done a performance study of iterative decoding algorithms. We have shown that by increasing the number of outer iterations error rate performance can be further improved. GPUs are specialized devices specially designed to accelerate graphics processing. They are massively parallel devices which can run thousands of threads simultaneously. Because of their tremendous processing power there is an increasing interest in using them for scientific and general purpose computations. Hence companies like Nvidia, Advanced Micro Devices (AMD) etc. have started their support for General Purpose GPU (GPGPU) applications. Nvidia came up with Compute Unified Device Architecture (CUDA) to program its GPUs. Efforts are made to come up with a standard language for parallel computing that can be used across platforms. OpenCL is the first such language which is supported by all major GPU and CPU vendors. MIMO detector has a high computational complexity. We have implemented a soft MIMO detector on GPUs and studied its throughput and latency performance. We have shown that a GPU can give throughput of up to 4Mbps for a soft detection algorithm which is more than sufficient for most general purpose tasks like voice communication etc. Compare to CPU a throughput increase of ~7x is achieved. We also compared the performances of two GPUs one with low computational power and one with high computational power. These comparisons show effect of thread serialization on algorithms with the lower end GPU's execution time curve shows a slope of 1/2. To further improve error rate performance iterative decoding techniques are employed where a feedback path is employed between detector and decoder. With an eye towards GPU implementation we have explored these algorithms. Better error rate performance however, comes at a price of higher power dissipation and more latency. By simulations we have shown that one can predict based on the Signal to Noise Ratio (SNR) values how many iterations need to be done before getting an acceptable Bit Error Rate (BER) and Frame Error Rate (FER) performance. Iterative decoding technique shows that a SNR gain of ~1:5dB is achieved when number of outer iterations is increased from zero. To reduce the complexity one can adjust number of possible candidates the algorithm can generate. We showed that where a candidate list of 128 is not sufficient for acceptable error rate performance for a 4x4 MIMO system using 16-QAM modulation scheme, performances are comparable with the list size of 512 and 1024 respectively.