Channel estimation is the process of understanding and analyzing the wireless communication channel's properties. It helps optimize data transmission by providing essential information for adjusting encoding and decoding parameters. This thesis explores using a Convolutional Neural Network~(CNN) for channel estimation in the 5G Link Level Simulator, 5G-LLS, developed by Tietoevry. The objectives were to create a Python framework for channel estimation experimentation and to evaluate CNN's performance compared to the conventional algorithms Least Squares~(LS), Minimum Mean Square Error~(MMSE) and Linear Minimum Mean Square Error~(LMMSE). Two distinct channel model scenarios were investigated in this study. The results from the study suggest that CNN outperforms LMMSE, LS, and MMSE regarding Mean Squared Error~(MSE) for both channel models, with LMMSE at second place. It managed to lower to the MSE by 85\% compared to the LMMSE for the correlated channel and 78\% for the flat fading channel. In terms of the overall system-level performance, as measured by Bit-Error Rate (BER), the CNN only managed to outperform LS and MMSE. The CNN and the LMMSE yielded similar results. This was due to that the LMMSE's MSE was still good enough to demodulate the symbols for the QPSK modulation scheme correctly.  The insights in this thesis work enables Tietoevry to implement more machine learning algorithms and further develop channel estimation in 5G telecommunications and wireless communication networks through experiments in 5G-LLS. Given that the CNN did not increase the performance of the communication system, future studies should test a broader range of channel models and consider more complex modulation schemes. Also, studying other and more advanced machine learning techniques than CNN is an avenue for future research. / Kanalestimering är en process i trådlösa kommunikationssystem som handlar om att analysera och förstå det trådlösa mediumets egenskaper. Genom effektiv kanalestimering kan dataöverföringen optimeras genom att anpassa signalen efter den trådlösa kanalen. Detta arbete utforskar användningen av ett konvolutionellt neuralt nätverk (CNN) för kanalestimering i Tietoevrys 5G-datalänkslagersimulator (5G-LLS). Målen är att (1) skapa ett Python-ramverk för kanalestimeringsexperiment samt att (2) utvärdera CNN:s prestanda jämfört med konventionella algoritmerna minsta kvadratmetoden (LS), minimalt medelkvadratsfel (MMSE) och linjärt minimalt medelkvadratsfel (LMMSE). Två olika kanalmodellsituationer undersöks i detta arbete. Resultaten visar att CNN överträffar LMMSE, LS och MMSE i form av medelkvadratisk fel (MSE) för båda kanalmodellerna, med LMMSE på andra plats. CNN:n lyckades minska MSE:n med 85\% jämfört med LMMSE för den korrelerade kanalen och med 78\% för den snabbt dämpande kanalen. Vad gäller systemnivåprestanda, mätt med hjälp av bitfelsfrekvens (BER), lyckades CNN endast överträffa LS och MMSE. CNN och LMMSE gav liknande resultat. Detta beror på att LMMSE:s MSE fortfarande var tillräckligt låg för att korrekt demodulera symbolerna för QPSK-modulationsschemat. Resultatet från detta examensarbete möjliggör för Tietoevry att implementera fler maskininlärningsalgoritmer och vidareutveckla kanalestimering inom 5G-telekommunikation och trådlösa kommunikationsnätverk genom experiment i 5G-LLS. Med tanke på att CNN inte överträffade samtliga kanalestimeringstekniker bör framtida studier testa ett bredare utbud av kanalmodeller och överväga mer komplexa moduleringsscheman. Framtida arbeten bör även utforska fler och mer avancerade maskininlärningsalgoritmer än CNN.

Machine learning (ML) has shown its success in many areas such as computer vision, natural language processing, robot control, and gaming. ML also draws significant attention in the wireless communication society. However, applying ML schemes to wireless communication networks is not straightforward, there are several challenges need to addressed: 1). Training data in communication networks, especially in physical and MAC layer, are extremely limited; 2). The high-dynamic wireless environment and fast changing transmission schemes in communication networks make offline training impractical; 3). ML tools are treated as black boxes, which lack of explainability. This dissertation tries to address those challenges by selecting training-efficient neural networks, devising online training frameworks for wireless communication scenarios, and incorporating communication domain knowledge into the algorithm design. Training-efficient ML algorithms are customized for three communication applications: 1). Symbol detection, where real-time online learning-based symbol detection algorithms are designed for MIMO-OFDM and massive MIMO-OFDM systems by utilizing reservoir computing, extreme learning machine, multi-mode reservoir computing, and StructNet; 2) Channel estimation, where residual learning-based offline method is introduced for WiFi-OFDM systems, and a StructNet-based online method is devised for MIMO-OFDM systems; 3) Radio resource management, where reinforcement learning-based schemes are designed for dynamic spectrum access, as well as ORAN intelligent network slicing management. All algorithms introduced in this dissertation have demonstrated outstanding performance in their application scenarios, which paves the path for adopting ML-based solutions in practical wireless networks. / Doctor of Philosophy / Machine learning (ML), which is a branch of computer science that trains machine how to learn a solution from data, has shown its success in many areas such as computer vision, natural language processing, robot control, and gaming. ML also draws significant attention in the wireless communication society. However, applying ML schemes to wireless communication networks is not straightforward, there are several challenges need to addressed: 1). Training issue: unlike areas such as computer vision where large amount of training data are available, the training data in communication systems are limited; 2). Uncertainty in generalization: ML usually requires offline training, where the ML models are trained by artificially generated offline data, with the assumption that offline training data have the same statistical property as the online testing one. However, when they are statistically different, the testing performance can not be guaranteed; 3). Lack of explainability, usually ML tools are treated as black boxes, whose behaviors can hardly be explained in an analytical way. When designed for wireless networks, it is desirable for ML to have similar levels of explainability as conventional methods. This dissertation tries to address those challenges by selecting training-efficient neural networks, devising online training frameworks for wireless communication scenarios, and incorporating communication domain knowledge into the algorithm design. Training-efficient ML algorithms are customized for three communication applications: 1). Symbol detection, which is a critical step of wireless communication receiver processing, it aims to recover the transmitted signals from the corruption of undesired wireless channel effects and hardware impairments; 2) Channel estimation, where transmitter transmits a special type of symbol called pilot whose value and position are known for the receiver, receiver estimates the underlying wireless channel by comparing the received symbols with the known pilots information; 3) Radio resource management, which allocates wireless resources such bandwidth and time slots to different users. All algorithms introduced in this dissertation have demonstrated outstanding performance in their application scenarios, which paves the path for adopting ML-based solutions in practical wireless networks.

No description available.

No description available.

Multiple Input Multiple Output (MIMO) and Orthogonal Frequency Division Multiplexing (OFDM) are two efficient technologies used to achieve higher data rate, lowlatency, robustness against fading used in 5G New Radio (NR). At the receiver end,the data arrives distorted due to disturbance during transfer over the wireless channel.Channel estimation is the applied technique at the receiver end to overcome this problemand mitigate the effect of the disturbance over the wireless channel. The main objective of this thesis is to port an existing channel estimator from a prototypesimulator for 5G to a complete Link-Level simulator that currently has support for 4Gtraffic. Two channel estimation algorithms have been investigated and implemented inthe Link-Level simulator based on MIMO-OFDM system. The channel estimators arethe Least Square (LS) and the Linear Minimum Mean Square Error (LMMSE). Theperformance of the channel estimators is evaluated in terms of Bit Error Rate vs Signalto Noise Ratio. The effectiveness of those implemented algorithms is evaluated using a simulation,where the results show that each channel estimation algorithm is suitable for a specificuse case and depends on channel properties and different scenarios but regardless thetime complexity, the LMMSE has better performance than the LS.

The IEEE 802.11 standards define several transmission rates that can be used at the physical layer to adapt the transmission rate to channel conditions. This dynamic adaptation attempts to improve the performance in Wireless LAN (WLAN) and hence can have impact on the Quality of Service (QoS) perceived by the users. In this work we present the design and implementation of several new link adaptation (LA) algorithms. The performance of the developed algorithms is tested and compared against some existing algorithms such as Minstrel as well as an ideal LA. The evaluation is carried out in a network system simulator that models all the pro- cedures needed for the exchange of data frames according to the 802.11 standards. Different scenarios are used to simulate various realistic conditions. In particular, the Clear Channel Assessment Threshold (CCAT) is modified in the scenarios and the impact of its modification is also assessed. The algorithms are tested under identical environments to ensure that the experiments are controllable and repeatable. For each algorithm the mean and 5th percentile throughput are measured under different traffic loads to evaluate and compare the performance of the different algorithms. The tradeoff between signaling overhead and performance is also evaluated. It was found that the proposed link adaptation schemes achieved higher mean through- put than the Minstrel algorithm. We also found that the performance of some of the proposed schemes is close to that of the ideal LA. / IEEE 802.11-standarderna definierar flera överföringshastigheter som kan användas vid det fysiska skiktet för att anpassa överföringshastigheten till kanal förhållanden. Denna dynamiska anpassning försöker förbättra prestandan i wireless LAN (WLAN) och därmed kan ha inverkan på Quality of Service (QoS) uppfattas av användarna. I detta examensarbete presenterar vi utformningen och genomförandet av flera ny link adaptation (LA) algoritmer. Prestandan hos de utvecklade algoritmer testas och jämförs med vissa befintliga algoritmer så som Minstrel liksom en ideal LA. Utvärderingen genomförs i ett nätverkssystem simulator som ger alla de förfaranden som behövs för utbyte av dataramar enligt 802.11-standarderna. Olika scenarier används för att simulera olika realistiska förhå llanden. Algoritmerna är testade under identiska miljöer för att experimenten är styrbar och repeterbar. För varje algoritm genomströmningen mättes under olika trafikbelastningar för att utvärdera och jämföra resultaten för de olika algoritmer. Den avvägning mellan signalering overhead och prestanda utvärderas också . Det konstaterades att de system som föreslå s link adaptation uppnå s högre genom- snittlig throughput än Minstrel algoritm. Vi fann också att utförandet av vissa av de föreslagna systemen är nära den av ideal LA.

No / Two training schemes namely the orthogonal pilot scheme (OPS) and the superimposed pilot scheme (SPS) are compared in terms of achievable rates in multiple-antenna fading channels with memory. For both schemes, we show that the achievable rate depends on the number of antennas, signal-to-noise ratio (SNR) and fading speed via the channel estimation error variance and the fraction of time for data transmission. To guarantee positive achievable rates, we show that for the OPS the number of transmit antennas that can be accommodated is limited by the fading speed whereas for the SPS the number of antennas can be arbitrary. For most antenna configurations, we observe that while the SPS is superior in the low-SNR and fastfading regimes, the OPS is superior in other regimes. However, for a few number of antennas (e.g., single antenna), the SPS may also be superior in the low-SNR and slow-fading regimes.

Space Time Trellis Coding (STTC) is a unique technique that combines the use of multiple transmit antennas with channel coding. This scheme provides capacity benefits in fading channels, and helps in improving the data rate and reliability of wireless communication. STTC schemes have been primarily designed assuming perfect channel estimates to be available at the receiver. However, in practical wireless systems, this is never the case. The noisy wireless channel precludes an exact characterization of channel coefficients. Even near-perfect channel estimates can necessitate huge overhead in terms of processing or spectral efficiency. This practical concern motivates the study of the impact of channel estimation errors on the design and performance of STTC. The design criteria for STTC are validated in the absence of perfect channel estimates at the receiver. Analytical results are presented that model the performance of STTC systems in the presence of channel estimation errors. Training based channel estimation schemes are the most popular choice for STTC systems. The amount of training however, increases with the number of transmit antennas used, the number of multi-path components in the channel and a decrease in the channel coherence time. This dependence is shown to decrease the performance gain obtained when increasing the number of transmit antennas in STTC systems, especially in channels with a large Doppler spread (low channel coherence time). In frequency selective channels, the training overhead associated with increasing the number of antennas can be so large that no benefit is shown to be obtained by using STTC. The amount of performance degradation due to channel estimation errors is shown to be influenced by system parameters such as the specific STTC code employed and the number of transmit and receive antennas in the system in addition to the magnitude of the estimation error. Hence inappropriate choice of system parameters is shown to significantly alter the performance pattern of STTC. The viability of STTC in practical wireless systems is thus addressed and it is shown that that channel estimation could offset benefits derived from this scheme. / Master of Science

<p dir="ltr">Wireless technologies across various applications aim to improve further by developing intelligent systems, where the performance is optimized through adaptive policy selections that efficiently adjust to the environment dynamics. As a result, accurate observation on the surrounding conditions, such as wireless channel quality and relative target location, becomes an important task. Although both channel estimation and wireless positioning problems have been well studied, with advanced wireless communications relying on complex technologies and being applied to diverse environments, optimization strategies tailored to their unique architectures and scenarios need to be further investigated. In this dissertation, four key research problems related to channel estimation and wireless positioning tasks for intelligent wireless systems are identified and studied. First, a channel denoising problem in multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) systems is addressed, and a Q-learning-based successive denoising scheme, which utilizes a channel curvature magnitude threshold to recover unreliable channel estimates, is proposed. Second, a pilot assignment problem in scalable open radio access network (O-RAN) cell-free massive MIMO (CFmMIMO) systems is studied, where a low-complexity pilot assignment scheme based on a multi-agent deep reinforcement learning (MA-DRL) framework along with a codebook search strategy is proposed. Third, sensor selection/placement problems for wireless positioning are addressed, and dynamic and robust sensor selection schemes that minimize the Cramér-Rao lower bound (CRLB) are proposed. Lastly, a feature selection problem for deep learning-based wireless positioning is studied, and a unique feature size selection method, which weights over the expected information gain and classification capability, along with a multi-channel positioning neural network is proposed.</p>

The following thesis consists of two main contributions to the fields of network coding and resource allocation. The first is a quantitative analysis of the effects of channel estimation errors and time synchronization errors on the performance of different network coding algorithms. It is shown that the performance improvement gained by a relay-based network scheme is significant for small number of users and when the quality of the relay link is better than that of the direct link. However, it is shown that potential performance improvement resulting from the considered relay-based network coding scheme could be negated by channel estimation errors. To consider the effects of time synchronization errors, we study a digital network coding (DNC) system and a physical-layer network coding (PNC) system with non-coherent frequency shift keying (FSK) modulation. For each of these two systems, we investigate the effects of received Eb/N0, unequal link quality, and time synchronization errors. The second contribution is an analysis of the value and cost of cognition obtained by investigating three resource allocation algorithms with different levels of channel knowledge in the context of ad hoc networks. The performance (quantified in terms of "percentage of users reaching target data rate" and "average effective data rate") and cost ("power consumption" and "number of channel estimations") of these algorithms are analyzed. Results show that a resource allocation algorithm with a higher level of channel knowledge results in better performance, but greater cost in terms of number of channel estimations, as expected. In addition, a resource allocation algorithm with a higher level of channel knowledge converges quicker when channel adaptation are necessary. Both an ideal medium access control (MAC) protocol and a non-ideal MAC protocol (dedicated control channel) are considered. / Master of Science