Over the past three decades, agricultural machinery has made the transition from purely mechanical systems to hybrid machines, reliant on both mechanical and electronic systems. A this transformation continues, the most modern agricultural machinery uses networked systems that require a network connection to function to their full potential. In rural areas, providing this network connection has proven difficult. Obstacles, distance from access points, and incomplete coverage of cellular connection are all challenges to be overcome. “Off the shelf” commercial-grade Wi-Fi equipment, including many products from Ubiquiti like the Bullet M2 transceiver and the PowerBeam point-to-point linking system, as well as antennas by Terrawave, Crane, and Hawking, were installed in a purpose-built system which could be implemented on a production farm. This system consisted of a tower-mounted access point which used an antenna with a 65^o beamwidth, and the test included distances up to 1150 meters in an agricultural setting with corn and soybeans. Some sensors were stationary and the other platform was a tractor following a path around the farm with both 8dBi and 15dBi gain antennas. Through all tests, throughput never dropped below 5 Mb/s, and the latency of successful connections never exceeded 20ms. Packets were rarely dropped and never accounted for a significant portion of all packet transmission attempts. Environmental effects like immediate precipitation, crop heights, recent rainfall, and ambient temperature had little or no effect on wireless network characteristics. As a result, it was proven that as long as line-of-sight was maintained, reliable wireless connectivity could be achieved despite varying conditions using microwave radiation. Network throughput was marginally affected by the change in free space path loss due to increased distance between the access point and the client, as well as travel by the mobile client outside the beamwidth of the access point. By enabling this coverage, it is hoped that the implementation of new agricultural technology utilizing a live network connection will progress more rapidly.

Future high-capacity wireless communications will extensively use the broad bands still available millimeter-wave frequencies. Channels with bandwidth broader than those in use today will guarantee enhanced data-rate and reduced latency performance. The recent progress of integrated-circuit semiconductor technologies finally allowed the design of reliable electronics operating at millimeter-wave frequencies. On top, advanced Fully Depleted Silicon On Insulator (FD-SOI) Complementary Metal Oxide Semiconductor (CMOS) and Silicon Germanium (SiGe) Bipolar CMOS (BiCMOS) processes enabled to co-integrate large digital blocks with frontends operating at tens or hundreds of GHz. The current under-deployment fifth-generation mobile-communication standard (5G) takes advantage of these advancements, massively exploiting the frequency bands from 24 GHz to 100 GHz. Furthermore, besides enlarging the channel bandwidth, improvements of the signal-to-noise power ratio (SNR) at the receiver input, combined with Multiple-Input Multiple-Output (MIMO) techniques provide an additional boost to the communication data-rate. Both approaches require arrays of antennas, plus electronic beam-steering which becomes essential in the case of moving transmitting-receiving pairs. Finally, social, economic, historical, and technological trends indicate that future wireless standards will require data-rates, latencies, and density of served users per square kilometer well beyond those offered by the 5G. Envisioned to be deployed towards the end of this decade, the six mobile communication standard (6G) will win future challenges thanks to the very ultra-broad bands available from 100 GHz until the tens of THz. Basic research is hence needed to address the open challenges necessary to reach the goals of future wireless communication systems, such as bandwidth and frequency operation factor-10 increase or power consumption reduction against the actual state of the art. This Habilitation thesis proposes circuit theory and concepts up to feasibility study of circuit implementation and experimental characterization in the laboratory of transceiver electronics for future high-capacity communications useful for the knowledge gain necessary for the conception of future communication systems. In detail, basic scientific research to understand the operation of millimeter-wave communication circuits implemented in 22 nm FD-SOI CMOS and 130 nm SiGe BiCMOS technologies has been performed.

Les stations de base à petite échelle (picocellules et femtocellules) seront un des leviers principaux qui permettront d'atteindre l'objectif 1000X, objectif fixé par les grands acteurs du domaine des télécommunications visant à augmenter la capacité des réseaux mobiles sans fil 5G d'un facteur 1000 par rapport aux réseaux 4G. Dans ce type de réseau, l'amplificateur de puissance (PA) est responsable de la majorité de la consommation de puissance de la station de base. Pour minimiser sa consommation de puissance, le PA est polarisé proche de sont point de compression mais avec l'augmentation des largeurs de bande, ce dernier subit des effets de mémoire accrus qui viennent s'ajouter aux problèmes classiques de non-linéarités. Les systèmes de prédistorsion numérique (DPD), et analogique/RF(ARFPD) peuvent être utilisés pour améliorer le compromis linéarité / efficacité des PAs. Cependant pour les pico-cellules et femto-cellules utilisées dans le standard 5G, les prédistorseurs conventionnels ne sont adaptés pour des raisons de complexité et de consommation de puissance.Le modèle "Memory Polynomilal" (MP) est l'un des modèles de prédistorsion les plus attractifs pour modéliser les PAs, fournissant des performances intéressantes avec peu de coefficients. Cependant, la précision de ce modèle se dégrade pour les signaux large bance. Pour palier ce problème, nous proposons un nouveau modèle, le FIR-MP qui combine un filtre FIR au modèle MP classique. Pour valider et quantifier la précision du modèle proposé, nous avons effectué des simulations avec un modèle extrait par mesure de l'amplificateur sur étagère ADL5606 (GaAs 1W HBT PA). Les résultats de ces simulations présentent des améliorations du taux de fuite des canaux adjacents (ACLR) de 7,2 dB et 15,6 dB, respectivement, pour des signaux à 20 MHz et 80 MHz par rapport au modèle MP classique. Le FIR-MP a été également synthétisé en technologie CMOS FDSOI 28 nm. Les résultats de la synthèse ont donné une puissance globale de 9,18 mW and 116,2 mW, respectivement, pour les signaux de 20 MHz and 80 MHz.Basé sur le modèle proposé de FIR-MP, une nouvelle approche à signaux mixtes pour linéariser les PAs a été aussi étudiée. En fait, le filtre numérique FIR améliore la performance de correction de la mémoire sans aucune expansion de la bande passante et la linéarisation en bande de base permet d'éviter l'utilisation de composants RF dans la linéariseur. Ainsi, les contraintes en bande passante requises pour le DAC, les filtres de reconstruction et les blocs RF de l'émetteur sont relâchées comparés aux techniques conventionnelles de linéarisation numériques et RF. Nous avons ainsi étudié l'impact des diverses non-idéalités en utilisant un signal modulé à 80 MHz afin de dériver les exigences pour la mise en œuvre du circuit. Les simulations ont montré qu'une résolution de 8 bits pour les coefficients et un SNR de 60 dB sont nécessaires pour atteindre un ACLR1 supérieur à 45 dBc. Ces résultats constituent un premier signe favorable dans l'optique d'une implémentation matérielle de la solution proposée, étape indispensable pour évaluer précisément sa consommation de puissance et sa complexité pour pouvoir la comparer à l'état de l'art des linéariseurs. / Small-cell base stations (picocells and femtocells) handling high bandwidths (> 100 MHz) will play a vital role in realizing the 1000X network capacity objective of the future 5G wireless networks. Power Amplifier (PA) consumes the majority of the base station power, whose linearity comes at the cost of efficiency. With the increase in bandwidths, PA also suffers from increased memory effects. Digital predistortion (DPD) and analog RF predistortion (ARFPD) tries to solve the linearity/efficiency trade-off. In the context of 5G small-cell base stations, the use of conventional predistorters becomes prohibitively power-hungry.Memory polynomial (MP) model is one of the most attractive predistortion models, providing significant performance with very few coefficients. We propose a novel FIR memory polynomial (FIR-MP) model which significantly augments the performance of the conventional memory polynomial predistorter. Simulations with models extracted on ADL5606 which is a 1 W GaAs HBT PA show improvements in adjacent channel leakage ratio (ACLR) of 7.2 dB and 15.6 dB, respectively, for 20 MHz and 80 MHz signals, in comparison with MP predistorter. Digital implementation of the proposed FIR-MP model has been carried out in 28 nm FDSOI CMOS technology. With a fraction of the power and die area of that of the MP a huge improvement in ACLR is attained.An overall estimated power consumption of 9.18 mW and 116.2 mW, respectively, for 20 MHz and 80 MHz signals is obtained.Based on the proposed FIR-MP model a novel low-power mixed-signal approach to linearize RF power amplifiers (PAs) is presented. The digital FIR filter improves the memory correction performance without any bandwidth expansion and the MP predistorter in analog baseband provides superior linearization. MSPD avoids 5X bandwidth requirement for the DAC and reconstruction filters of the transmitter and the power-hungry RF components when compared to DPD and ARFPD, respectively.The impact of various non-idealities is simulated with ADL5606 (1 W GaAs HBT PA) MP PA model using 80 MHz modulated signal to derive the requirements for the integrated circuit implementation. A resolution of 8 bits for the coefficients and a signal path SNR of 60 dB is required to achieve ACLR1 above 45 dBc, with as little as 9 coefficients in the analog domain. Discussion on the potential circuit architectures of subsystems is provided. It results that an analog implementation is feasible. It will be worth in the future to continue the design of this architecture up to a silicon prototype to evaluate its performance and power consumption.

Med framväxten av framtida mobilgenerationer bortom 5G, studeras nya teknologier för att tillgodose de förväntade kraven för framtida tjänster som Ultra-Reliable Low Latency Communications (URLLC) eller Virtual Reality. Bland dessa teknologier uppstår Reconfigurable Intelligent Surfaces (RIS) som en av de mest lovande på grund av deras förmåga att förbättra kanalen samtidigt som de bara ökar nätverkets energiförbrukning måttligt. Men flera utmaningar måste lösas innan de kan distribueras. I denna avhandling studerar vi strategier för att positionera RIS för att uppnå maximal SNR-täckning i en utomhusförökningsmiljö. Vår modell tar hänsyn till effekterna av skuggfading och siktlinje (LoS). En jämförelse mellan centraliserade och distribuerade distributioner övervägs också. Dessutom bedöms den nödvändiga storleken på RIS för att matcha täckningen av en liten cell. Resultaten visar att de bästa positionerna för att distribuera en RIS ligger nära de mobila terminalerna, i närheten av gränsen mellan täckta och utomtäckta områden. Man drar slutsatsen att en centraliserad distribution är bättre än en distribuerad, och en genomförbar storlek på RIS som matchar den lilla celltäckningen erhålls. / With the emergence of future mobile generations beyond 5G, novel technologies are studied to satisfy the envisioned requirements of future services such as Ultra-Reliable Low Latency Communications (URLLC) or Virtual Reality. Among these technologies, Reconfigurable Intelligent Surfaces (RIS) arise as one of the most promising due to their capabilities to improve the channel while only modestly increasing the network energy consumption. However, multiple challenges have to be addressed before they can be deployed. In this thesis, we study strategies for positioning the RIS to achieve maximum SNR coverage in an outdoor propagation environment. Our model takes into account the effects of shadow fading and line-of-sight (LoS). A comparison between centralized and distributed deployments is also considered. Additionally, the required size of RIS to match the coverage of a small cell is assessed. The results show that the best positions to deploy a RIS lie close to the mobile terminals, in the vicinity of the boundary between covered and out-of-coverage areas. It is concluded that a centralized deployment is better than a distributed one, and a feasible size of the RIS which matches the small cell coverage is obtained.

With the ever-growing demand for higher data rate in next generation communication systems, researchers are pushing the limits of the existing architecture. Due to the stochastic nature of communication channels, most systems use some form of adaptive methods to adjust the transmitting parameters and allocation of resources in order to overcome channel variations and achieve optimal throughput. We will study four cases of adaptive transmission and dynamic resource allocation in collaborative systems that are practically significant. Firstly, we study hybrid automatic repeat request (HARQ) techniques that are widely used to handle transmission failures. We propose HARQ policies that improve system throughput and are suitable for point-to-point, two-hop relay, and multi-user broadcast systems. Secondly, we study the effect of having bits of mixed SNR qualities in finite length codewords. We prove that by grouping bits according to their reliability so that each codeword contains homogeneous bit qualities, the finite blocklength capacity of the system is increased. Thirdly, we study the routing and resource allocation problem in multiple collaborative networks. We propose an algorithm that enables collaboration between networks which needs little to no side information shared across networks, but rather infers necessary information from the transmissions. The collaboration between networks provides a significant gain in overall throughput compared to selfish networks. Lastly, we present an algorithm that allocates disjoint transmission channels for our cognitive radio network in the DARPA Spectrum Collaboration Challenge (SC2). This algorithm uses the real-time spectrogram knowledge perceived by the radios and allocates channels adaptively in a crowded spectrum shared with other collaborative networks.

This dissertation provides the theory, design, and implementation of a cognitive engine, the enabling technology of cognitive radio. A cognitive radio is a wireless communications device capable of sensing the environment and making decisions on how to use the available radio resources to enable communications with a certain quality of service. The cognitive engine, the intelligent system behind the cognitive radio, combines sensing, learning, and optimization algorithms to control and adapt the radio system from the physical layer and up the communication stack. The cognitive engine presented here provides a general framework to build and test cognitive engine algorithms and components such as sensing technology, optimization routines, and learning algorithms. The cognitive engine platform allows easy development of new components and algorithms to enhance the cognitive radio capabilities. It is shown in this dissertation that the platform can easily be used on a simulation system and then moved to a real radio system. The dissertation includes discussions of both theory and implementation of the cognitive engine. The need for and implementation of all of the cognitive components is strongly featured as well as the specific issues related to the development of algorithms for cognitive radio behavior. The discussion of the theory focuses largely on developing the optimization space to intelligently and successfully design waveforms for particular quality of service needs under given environmental conditions. The analysis develops the problem into a multi-objective optimization process to optimize and trade-of of services between objectives that measure performance, such as bit error rate, data rate, and power consumption. The discussion of the multi-objective optimization provides the foundation for the analysis of radio systems in this respect, and through this, methods and considerations for future developments. The theoretical work also investigates the use of learning to enhance the cognitive engine's capabilities through feed-back, learning, and knowledge representation. The results of this work include the analysis of cognitive radio design and implementation and the functional cognitive engine that is shown to work in both simulation and on-line experiments. Throughout, examples and explanations of building and interfacing cognitive components to the cognitive engine enable the use and extension of the cognitive engine for future work. / Ph. D.

<p>Deep learning has been increasingly adopted in a multitude of settings. Yet, its strong performance relies on processing data during inference that is in-distribution with its training data. Deep learning input data during deployment, however, is not guaranteed to be in-distribution with the model's training data and can often times be distorted, either intentionally (e.g., by an adversary) or unintentionally (e.g., by a sensor defect), leading to significant performance degradations. In this dissertation, we develop algorithms for a variety of applications to improve the performance of deep learning models in the presence of distorted data. We begin by first designing feature engineering methodologies to increase classification performance in noisy environments. Here, we demonstrate the efficacy of our proposed algorithms on two target detection tasks and show that our framework outperforms a variety of state-of-the-art baselines. Next, we develop mitigation algorithms to improve the performance of deep learning in the presence of adversarial attacks and nonlinear signal distortions. In this context, we demonstrate the effectiveness of our methods on a variety of wireless communications tasks including automatic modulation classification, power allocation in massive MIMO networks, and signal detection. Finally, we develop an uncertainty quantification framework, which produces distributive estimates, as opposed to point predictions, from deep learning models in order to characterize samples with uncertain predictions as well as samples that are out-of-distribution from the model's training data. Our uncertainty quantification framework is carried out on a hyperspectral image target detection task as well as on counter unmanned aircraft systems (cUAS) model. Ultimately, our proposed algorithms improve the performance of deep learning in several environments in which the data during inference has been distorted to be out-of-distribution from the training data. </p>

The industrial needs in the production of goods and control of processes within the factory keep leapfrogging daily by the necessities to fulfil the needs of the ever-growing population. In recent times, the industries are looking towards Industry 4.0 to improve their overall productivity and scalability. One of the significant aspects that are required to meet the requirements of Industry 4.0 is communication networks among industrial applications. Nowadays, industries from the cross markets are looking to replace their existing wired networks with wireless networks, which indeed brings many use-cases and a lot of new business models into existence. To make all these options possible, wireless networks need to meet the stringent requirements of these industrial applications in the form of reliability, latency, and service availability. This thesis focuses on a systematic methodology to integrate wireless networks like 5G, Wi-Fi 6, etc., into real-life automation devices. It also describes a methodology to evaluate their communication and control performance by varying control parameters like topology, cycle time, and type of networks. It also devises some techniques and methods that can improve the overall performance, i.e., both control and communication performance of the control applications. The method used to implement this work is a case study. This work integrates and tests the industrial applications in a real-life scenario. It is the best effort to bring a unique perspective of communication engineers and control engineers together regarding the performance of the industrial applications. This work tries to verify the suitability of the wireless in mission-critical control application scenarios with respect to their communication and control performance. Software for data analysis and visualization and its methodology for analyzing the traffic flow of the control applications via different wireless networks is demonstrated by varying different control parameters. It is shown that it is challenging for 5G to support the shorter cycle time values, and performance will get better and more stable with the increase in the cycle time of the control application. It is also found that the 1-Hop wireless topologies have a comparatively better control performance than 2-Hop wireless topologies. In the end, it is found that the communication and control performance of the motion control application can be improved by using the hybrid topology, which is a mixture of 5G and Wi-Fi 6, by modifying some key aspects. The thesis work helps to introduce a novel systematic methodology for measuring and analyzing the communication and control applications via different wireless networks. It also gives a better idea for the control engineers in the industry about which cycle times the different wireless networks and their topologies support when integrated with industrial automation devices. It also describes which wireless networks support industrial applications better. It ends with a novel methodology that could improve the performance of the mission-critical motion applications by using existing wireless technologies.

The ever-expanding volume of data generated by network devices such as smartphones, personal computers, and sensors has significantly contributed to the remarkable advancements in artificial intelligence (AI) and machine learning (ML) algorithms. However, effectively processing and learning from this extensive data usually requires substantial computational capabilities centralized in a server. Moreover, concerns regarding data privacy arise when collecting training data from distributed network devices. To address these challenges, collaborative ML with decentralized data has emerged as a promising solution for large-scale machine learning across distributed devices, driven by the parallel computing and learning trends. Collaborative and distributed ML can be broadly classified into two types: server-based and fully decentralized, based on whether the model aggregation is coordinated by a parameter server or performed in a decentralized manner through peer-to-peer communication. In cases where communication between devices occurs over wireless links, which are inherently imperfect, unreliable, and resource-constrained, how can we design communication protocols to achieve the best learning performance? This thesis investigates decentralized learning using decentralized stochastic gradient descent, an established algorithm for decentralized ML, in a novel setting with imperfect and constrained communication. "Imperfect" implies that communication can fail and "constrained" implies that communication resources are limited. The communication across a link between two devices is modeled as a binary event with either success or failure, depending on if multiple neighbouring devices are transmitting information. To compensate for communication failures, every communication round can have multiple communication slots, which are limited and must be carefully allocated over the learning process. The quality of communication is quantified by introducing normalized throughput, describing the ratio of successful links in a communication round. To decide when devices should broadcast, both random and deterministic medium access policies have been developed with the goal of maximizing throughput, which has shown very efficient learning performance. Finally, two schemes for allocating communication slots over communication rounds have been defined and simulated: Delayed-Allocation and the Periodic-Allocation schemes, showing that it is better to allocate slots late rather than early, and neither too frequently nor infrequently which can depend on several factors and requires further study

At present, the current 4G systems provide a universal platform for broadband mobile services; however, mobile traffic is still growing at an unprecedented rate and the need for more sophisticated broadband services is pushing the limits on current standards to provide even tighter integration between wireless technologies and higher speeds. This has led to the need for a new generation of mobile communications: the so-called 5G. Although 5G systems are not expected to penetrate the market until 2020, the evolution towards 5G is widely accepted to be the logical convergence of internet services with existing mobile networking standards leading to the commonly used term “mobile internet” over heterogeneous networks, with several Gbits/s data rate and very high connectivity speeds. Therefore, to support highly increasing traffic capacity and high data rates, the next generation mobile network (5G) should extend the range of frequency spectrum for mobile communication that is yet to be identified by the ITU-R. The mm-wave spectrum is the key enabling feature of the next-generation cellular system, for which the propagation channel models need to be predicted to enhance the design guidance and the practicality of the whole design transceiver system. The present work addresses the main concepts of the propagation channel behaviour using ray tracing software package for simulation and then results were tested and compared against practical analysis in a real-time environment. The characteristics of Indoor-Indoor (LOS and NLOS), and indoor-outdoor (NLOS) propagations channels are intensively investigated at four different frequencies; 5.8 GHz, 26GHz, 28GHz and 60GHz for vertical polarized directional, omnidirectional and isotropic antennas patterns. The computed data achieved from the 3-D Shooting and Bouncing Ray (SBR) Wireless Insite based on the effect of frequency dependent electrical properties of building materials. Ray tracing technique has been utilized to predict multipath propagation characteristics in mm-wave bands at different propagation environments. Finally, the received signal power and delay spread were computed for outdoor-outdoor complex propagation channel model at 26 GHz, 28 GHz and 60GHz frequencies and results were compared to the theoretical models.