Master of Science / Department of Electrical and Computer Engineering / Dwight D. Day / Low-power systems implemented on Field Programmable Gate Arrays (FPGA) have become more practical with advancements leading to decreases in FPGA power consumption, physical size, and cost. In systems that may need to operate for an extended time independent of a central power source, low-power FPGA’s are now a reasonable option. Combined with research into energy harvesting solutions, a FPGA-based system could operate independently indefinitely and be cost effective. Four simple demodulator designs were implemented on a FPGA to test and compare the performance and power consumption of each. The demodulators were a Counter that tracked the length of the input signal period, a One-Shot that counted the input edges over time, a Phase-Frequency Detector (PFD), and a PFD with preprocessing on the input signal to mitigate distortion introduces by the 1-bit subsampling. The designs demodulated a binary frequency shift keying (BFSK) signal using 10.69MHz and 10.71MHz as the input frequencies and a 1kHz data rate. The signal was 1-bit subsampled at 75kHz to provide the demodulators with a signal containing 15kHz and 35kHz. The design size, power consumption, and error performance of each demodulator were compared. At the frequencies and data rate used, the Counter and One-Shot are the most energy efficient by a significant margin over the PFDs. The error performance was nearly equal for all four. As the BFSK baseband frequencies and especially the data rate are increased, the PFD options are expected to be the better options as the Counter and One-Shot may not react quickly enough.

This thesis documents a research project in which an autonomous flight control system (FCS) was designed to control and navigate small, high-speed, unmanned, jet-turbine powered fixed-wing aircraft. The FCS was designed to allow the aircraft to maintain controlled flight, and return to a home location, without any operator intervention. The flight control computer was built with an FPGA, using a Microblaze soft-core microprocessor running the uClinux operating system. The configurable FPGA computing platform allowed flexibility for interfacing quickly with a wide range of sensors and control modules. A commercial inertial measurement unit was used for aircraft state estimation, and the flight control system was able to provide stability and precise flight-path control for multiple turbinepowered aircraft over the wide flight airspeed envelope these vehicles are capable of. In addition, the custom ground control station which provides an operator control interface for the FCS is discussed.

This work is part of ongoing research conducted at Virginia Commonwealth University relating to unmanned aerial vehicles. The primary objective of this thesis was to develop a flexible, high-performance autopilot platform in order to facilitate research on advanced flight control algorithms. A dual FPGA-based system architecture utilizing a stacked, multi-board design was created to meet this goal. Processing tasks were split between the two FPGA devices, allowing for improved system timing and increased throughput. A combination of analog and digital filtering techniques were employed in the new system, resulting in enhanced sensor accuracy and precision compared to the previous generation autopilot system. Several important improvements to the safety and reliability of the overall system were also achieved.

The aim of this thesis is to explore the possibility of integrating an AC3 audio de- coding module into the company’s current product. Due to limited left resources on the FPGA chip in the company’s current product, the focus of this thesis is to be resource efficient. In this thesis, a system for AC3 audio decoding is designed and implemented. In order to use less logic on FPGA, PicoBlaze soft processor is used to control the whole processing flow. The system is designed and synthe- sized for a Spartan-6 FPGA which can be easily ported to the company’s current platform.

Cette thèse s'inscrit dans le cadre de la conception d'architectures reconfigurables. Plus précisément, il concerne les architectures matérielles adaptatives, ces dernières pouvant être modifiées du point de vue de leurs caractéristiques matérielles au cours de l'exécution d'une application. Nous présentons une méthodologie d'auto-configuration d'une architecture reconfigurable dynamiquement ainsi qu'une architecture permettant d'illustrer l'utilisation de la méthode. L'objectif de la méthode est de réduire la consommation d'énergie en garantissant le respect des contraintes à tout instant. La méthodologie proposée s'adresse aux architectures reconfigurables à grain épais, puisque l'unité fonctionnelle matérielle correspond à une fonction de haut niveau d'abstraction (IDWT, etc.), même si la réalisation de l'architecture est basée sur l'utilisation d'une structure reconfigurable à grain fin (FPGA). Le besoin d'adaptation choisi concerne principalement deux cas de figures. Premièrement, répondre aux variations dynamiques de la charge de calcul en cours de traitement : un accroissement ou une réduction du débit de données conduit à une inadéquation entre l'architecture et son environnement. Deuxièmement, s'adapter aux variations dynamiques de la structure de l'algorithme : dans certaines applications les traitements à effectuer changent en fonction des données qui arrivent. / My PhD project focuses on Dynamic Adaptive Runtime parallelism and frequency scaling techniques in coarse grain reconfigurable hardware architectures. This new architectural approach offers a set of new features to increase the flexibility and scalability for applications in an evolving environment with reasonable energy cost. In this architecture, the parallelism granularity and running frequency can be reconfigured by using partial and dynamic reconfiguration. The adaptive method and architecture have been already developed and tested on FPGA platforms. The measurements and results analysis based on DWT show that the energy efficiency is adjustable dynamically by using our approach. The main contribution to the research project involves an auto-adaptive method development; this means using partial and dynamic reconfiguration can reconfigure the parallelism granularity and running frequency of application. The adaptive method by adjusting the parallelism granularity and running frequency is tested with the same application. We are presenting results coming from implementations of Image processing key application and analyses the behavior of this architecture on these applications.

Time to digital converter (TDC) is a digital unit that measures the time interval between two events.This is useful to determine the characteristics and patterns of a signal or an event. In this thesis ahybrid TDC is presented consisting of a tapped delay line and a clock counter principle. The TDC is used to measure the time between received data in a QKD application. If the measuredtime does not exceed a certain value then data had been sent without any interception. It is alsopossible to use TDCs in other fields such as laser-ranging and time-of-flight applications. The TDC consists of two carry chains, an encoder, a FIFO and a counter for each channel, anAXI-module and a control unit to generate command signals to all channels that are implemented.The time is measured by sampling the signal that has propagated through the carry chain and from thissample encode the propagation length. In this thesis a TDC is implemented that has a 10 ns dead time and a resolution below 28 psin a four channel mode. The propagation variation is approximately two percent of the total valueduring testing. For the implementation an FPGA-board with a Zynq XC7Z020 SoC is used withSystemVerilog that is a hardware describing language (HDL).

Les besoins de performance des systèmes sur puce embarqués augmentant sans cesse pour satisfaire des applications de plus en plus complexes, de nouvelles architectures de traitement et de nouveaux paradigmes de calcul sont apparus. L'intégration au sein d'une même puce électronique de plusieurs dizaines, voire centaines d'éléments de calcul a donné naissance aux systèmes sur puce multiprocesseur (MultiProcessor Systems on Chip - MPSoC). Cette évolution permet d'obtenir une puissance de traitement parallèle considérable. Actuellement, les performances de tels systèmes reposent sur le support de communication et d'échange des données entre les blocs de calcul intégrés. La problématique du support de communication est de fournir une bande passante et une adaptabilité élevées, afin de pouvoir bénéficier efficacement du parallélisme potentiel de la puissance de calcul disponible des MPSoC. C'est dans ce contexte du besoin primordial de flexibilité et de bande passante que sont apparus les réseaux embarqués sur puce (Network-on-Chip - NoC) dont l'objectif est de permettre l'interconnexion optimisée d'un grand nombre d'éléments de calcul au sein d'une même puce électronique, tout en assurant l'exigence d'un compromis entre les performances de communication et les ressources d'interconnexion. De plus, l'apparition de la technologie FPGA reconfigurable dynamiquement a ouvert de nouvelles approches permettant aux MPSoC d'adapter leurs constituants en cours de fonctionnement et de répondre aux besoins croissant d'adaptabilité, de flexibilité et de la diversité des ressources des systèmes embarqués. Étant donnée cette évolution de complexité des systèmes électroniques et la diminution de la finesse de gravure, et donc du nombre croissant de transistors au sein d'une même puce, la sensibilité des circuits face aux phénomènes générant des fautes n'a de cesse d'augmenter. Ainsi, dans le but d'obtenir des systèmes sur puces performants et fiables, des techniques de détection, de localisation et de correction d'erreurs doivent être proposées au sein des NoC reconfigurables ou adaptatifs, où la principale difficulté réside dans l'identification et la distinction entre des erreurs réelles et des fonctionnements variables ou adaptatifs des éléments constituants ces types de NoC C'est dans ce contexte que nous proposons de nouveaux mécanismes et solutions architecturales permettant de contrôler le fonctionnement d'un NoC adaptatif supportant les communications d'une structure MPSOC, et afin de d'identifier et localiser avec précision les éléments défaillants d'une telle structure dans le but de les corriger ou de les isoler pour prévenir toutes défaillances du système / The need of performance of embedded Syxtena-on-Chlps (Socs) are increasing constantly to meet the requirements of applications becoming more and more complexes, and new processing architectures and new computing paradigms have emerged. The integration within a single chip of dozens, or hundreds of computing and processing elements has given birth to Mukt1 Pmcesmr Systena-on-Chp (MPSoC) allowing to feature a high level of parallel processing. Nowaday s, the performance of these systems rely on the communication medium between the interconnected processing elements. The problematic of the communication medium to feature a high bandwidth and flexibility is primordial in order to efficiently use the parallel processing capacity of the MPSoC In this context, Network-on-Chlps (NoCs) are developed where the aim is to allow the interconnection of a large number of elements in the same device while maintaining a tradeoff between performance and logical resources. Moreover, the emergence of the partial reconfigurable FPGA technology allows to the MPSoC to adapt their elements during its operation in order to meet the system requirements. Given this increasing complexity of the electronic systems and the shrinking size of the devices, the sensibility of the chip against phenomena generating fault has increased. Thereby, to design efficient and reliable Socs, new error detection and localization techniques must be proposed for the dynamic NoCs where the main difficulty is the identification and the distinction between real errors and adaptive behavior of the NoCs. In this context, we present new mechanisms and architectural solutions allowing to check during the system operation the correctness of dynamic NoCs in order to locate and isolate efficiently the faulty components avoiding a failure of the system

Demands on modern computing are becoming more intensive. Keeping up with these demands has increasing complexity. Moore's Law is in decline. Increasing the number of cores on a device has diminishing returns. Specialised architectures provide more efficient and higher performing processors. However, it is not always practical to include every architecture on every device. Running non-native tasks on architectures often results in a drop in performance. This research examines the benefits and limitations of Field Programmable Gate Arrays - Systems on Chip (FPGA-SoC) devices to provide flexible hardware accelerators for heterogeneous architectures. A number of topics are covered, including hardware acceleration of floating-point mathematical functions, dynamic reconfiguration and high-level synthesis. A number of case studies are presented. Dynamic reconfiguration is used to change the configuration of the FPGA at runtime, allowing the hardware accelerators to be changed depending on the current processor tasks. Changing accelerators at runtime has limitations, such as data perturbation. Context switching techniques are applied to the hardware to prevent loss of data and enable de-fragmentation of the FPGA. High level synthesis techniques are used in conjunction with the presented hardware accelerators to synthesise high-level languages into hardware descriptions with optimisations. Techniques for runtime synthesis of hardware accelerators are presented. These can be combined with dynamic reconfiguration to configure FPGAs with appropriate hardware accelerators from a high-level language at runtime. The research demonstrates that FPGA-SoC devices have the potential for providing reconfigurable accelerators for processors in heterogeneous architectures. Metrics show that the FPGA configurations can perform better than other commercial processors. It was demonstrated that it is possible to context switch hardware at runtime, meaning the most can be made of the FPGA-SoC at all times, even as situations change. However, there are many limitations that still need to be overcome, such as management of the implemented hardware, synthesis of new hardware at runtime, reconfiguration times, interfacing of hardware with software and the design of hardware accelerators.

Intelligent vehicle technologies are growing rapidly that can enhance road safety, improve transport efficiency, and aid driver operations through sensors and intelligence. Advanced driver assistance system (ADAS) is a common platform of intelligent vehicle technologies. Many sensors like LiDAR, radar, cameras have been deployed on intelligent vehicles. Among these sensors, optical cameras are most widely used due to their low costs and easy installation. However, most computer vision algorithms are complicated and computationally slow, making them difficult to be deployed on power constraint systems. This dissertation investigates several mainstream ADAS applications, and proposes corresponding efficient digital circuits implementations for these applications. This dissertation presents three ways of software / hardware algorithm division for three ADAS applications: lane detection, traffic sign classification, and traffic light detection. Using FPGA to offload critical parts of the algorithm, the entire computer vision system is able to run in real time while maintaining a low power consumption and a high detection rate. Catching up with the advent of deep learning in the field of computer vision, we also present two deep learning based hardware implementations on application specific integrated circuits (ASIC) to achieve even lower power consumption and higher accuracy. The real time lane detection system is implemented on Xilinx Zynq platform, which has a dual core ARM processor and FPGA fabric. The Xilinx Zynq platform integrates the software programmability of an ARM processor with the hardware programmability of an FPGA. For the lane detection task, the FPGA handles the majority of the task: region-of-interest extraction, edge detection, image binarization, and hough transform. After then, the ARM processor takes in hough transform results and highlights lanes using the hough peaks algorithm. The entire system is able to process 1080P video stream at a constant speed of 69.4 frames per second, realizing real time capability. An efficient system-on-chip (SOC) design which classifies up to 48 traffic signs in real time is presented in this dissertation. The traditional histogram of oriented gradients (HoG) and support vector machine (SVM) are proven to be very effective on traffic sign classification with an average accuracy rate of 93.77%. For traffic sign classification, the biggest challenge comes from the low execution efficiency of the HoG on embedded processors. By dividing the HoG algorithm into three fully pipelined stages, as well as leveraging extra on-chip memory to store intermediate results, we successfully achieved a throughput of 115.7 frames per second at 1080P resolution. The proposed generic HoG hardware implementation could also be used as an individual IP core by other computer vision systems. A real time traffic signal detection system is implemented to present an efficient hardware implementation of the traditional grass-fire blob detection. The traditional grass-fire blob detection method iterates the input image multiple times to calculate connected blobs. In digital circuits, five extra on-chip block memories are utilized to save intermediate results. By using additional memories, all connected blob information could be obtained through one-pass image traverse. The proposed hardware friendly blob detection can run at 72.4 frames per second with 1080P video input. Applying HoG + SVM as feature extractor and classifier, 92.11% recall rate and 99.29% precision rate are obtained on red lights, and 94.44% recall rate and 98.27% precision rate on green lights. Nowadays, convolutional neural network (CNN) is revolutionizing computer vision due to learnable layer by layer feature extraction. However, when coming into inference, CNNs are usually slow to train and slow to execute. In this dissertation, we studied the implementation of principal component analysis based network (PCANet), which strikes a balance between algorithm robustness and computational complexity. Compared to a regular CNN, the PCANet only needs one iteration training, and typically at most has a few tens convolutions on a single layer. Compared to hand-crafted features extraction methods, the PCANet algorithm well reflects the variance in the training dataset and can better adapt to difficult conditions. The PCANet algorithm achieves accuracy rates of 96.8% and 93.1% on road marking detection and traffic light detection, respectively. Implementing in Synopsys 32nm process technology, the proposed chip can classify 724,743 32-by-32 image candidates in one second, with only 0.5 watt power consumption. In this dissertation, binary neural network (BNN) is adopted as a potential detector for intelligent vehicles. The BNN constrains all activations and weights to be +1 or -1. Compared to a CNN with the same network configuration, the BNN achieves 50 times better resource usage with only 1% - 2% accuracy loss. Taking car detection and pedestrian detection as examples, the BNN achieves an average accuracy rate of over 95%. Furthermore, a BNN accelerator implemented in Synopsys 32nm process technology is presented in our work. The elastic architecture of the BNN accelerator makes it able to process any number of convolutional layers with high throughput. The BNN accelerator only consumes 0.6 watt and doesn't rely on external memory for storage.

Digital communications devices designed with application-specific integrated circuit (ASIC) technology suffer from one very significant limitation�the integrated circuits are not programmable. Therefore, deploying a new algorithm or an updated standard requires new hardware. Field-programmable gate arrays (FPGAs) solve this problem by introducing what is essentially reconfigurable hardware. Thus, digital communications devices designed on FPGAs are capable of accommodating multiple communications protocols without the need to deploy new hardware, and can support new protocols in a matter of seconds. In addition, FPGAs provide a means to update systems that are physical difficult to access. For these reasons, FPGAs provide us with an ideal platform for implementing adaptive communications algorithms. This thesis focuses on using FPGAs to implement an adaptive digital communications system. Using the Universal Software Radio Peripheral (USRP) as a base, this thesis aims to create a highly-adaptive, plug and play software-defined radio (SDR) that fits CubeSat form-factor satellites. Such a radio platform would enable CubeSat engineers to develop new satellites faster and with lower costs. This thesis presents a new system, the COSMIAC CubeSat SDR, that adapts the USRP platform to better suit the space and power limitations of a CubeSat.