A. James Clark School of Engineering
Permanent URI for this communityhttp://hdl.handle.net/1903/1654
The collections in this community comprise faculty research works, as well as graduate theses and dissertations.
Browse
2 results
Search Results
Item DESIGN SPACE EXPLORATION FOR SIGNAL PROCESSING SYSTEMS USING LIGHTWEIGHT DATAFLOW GRAPHS(2018) Li, Lin; Bhattacharyya, Shuvra S; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)Digital signal processing (DSP) is widely used in many types of devices, including mobile phones, tablets, personal computers, and numerous forms of embedded systems. Implementation of modern DSP applications is very challenging in part due to the complex design spaces that are involved. These design spaces involve many kinds of configurable parameters associated with the signal processing algorithms that are used, as well as different ways of mapping the algorithms onto the targeted platforms. In this thesis, we develop new algorithms, software tools and design methodologies to systematically explore the complex design spaces that are involved in design and implementation of signal processing systems. To improve the efficiency of design space exploration, we develop and apply compact system level models, which are carefully formulated to concisely capture key properties of signal processing algorithms, target platforms, and algorithm-platform interactions. Throughout the thesis, we develop design methodologies and tools for integrating new compact system level models and design space exploration methods with lightweight dataflow (LWDF) techniques for design and implementation of signal processing systems. LWDF is a previously-introduced approach for integrating new forms of design space exploration and system-level optimization into design processes for DSP systems. LWDF provides a compact set of retargetable application programming interfaces (APIs) that facilitates the integration of dataflow-based models and methods. Dataflow provides an important formal foundation for advanced DSP system design, and the flexible support for dataflow in LWDF facilitates experimentation with and application of novel design methods that are founded in dataflow concepts. Our developed methodologies apply LWDF programming to facilitate their application to different types of platforms and their efficient integration with platform-based tools for hardware/software implementation. Additionally, we introduce novel extensions to LWDF to improve its utility for digital hardware design and adaptive signal processing implementation. To address the aforementioned challenges of design space exploration and system optimization, we present a systematic multiobjective optimization framework for dataflow-based architectures. This framework builds on the methodology of multiobjective evolutionary algorithms and derives key system parameters subject to time-varying and multidimensional constraints on system performance. We demonstrate the framework by applying LWDF techniques to develop a dataflow-based architecture that can be dynamically reconfigured to realize strategic configurations in the underlying parameter space based on changing operational requirements. Secondly, we apply Markov decision processes (MDPs) for design space exploration in adaptive embedded signal processing systems. We propose a framework, known as the Hierarchical MDP framework for Compact System-level Modeling (HMCSM), which embraces MDPs to enable autonomous adaptation of embedded signal processing under multidimensional constraints and optimization objectives. The framework integrates automated, MDP-based generation of optimal reconfiguration policies, dataflow-based application modeling, and implementation of embedded control software that carries out the generated reconfiguration policies. Third, we present a new methodology for design and implementation of signal processing systems that are targeted to system-on-chip (SoC) platforms. The methodology is centered on the use of LWDF concepts and methods for applying principles of dataflow design at different layers of abstraction. The development processes integrated in our approach are software implementation, hardware implementation, hardware-software co-design, and optimized application mapping. The proposed methodology facilitates development and integration of signal processing hardware and software modules that involve heterogeneous programming languages and platforms. Through three case studies involving complex applications, we demonstrate the effectiveness of the proposed contributions for compact system level design and design space exploration: a digital predistortion (DPD) system, a reconfigurable channelizer for wireless communication, and a deep neural network (DNN) for vehicle classification.Item BASEBAND RADIO MODEM DESIGN USING GRAPHICS PROCESSING UNITS(2015) KIM, SCOTT C.; Bhattacharyya, Shuvra S; Electrical Engineering; Digital Repository at the University of Maryland; University of Maryland (College Park, Md.)A modern radio or wireless communications transceiver is programmed via software and firmware to change its functionalities at the baseband. However, the actual implementation of the radio circuits relies on dedicated hardware, and the design and implementation of such devices are time consuming and challenging. Due to the need for real-time operation, dedicated hardware is preferred in order to meet stringent requirements on throughput and latency. With increasing need for higher throughput and shorter latency, while supporting increasing bandwidth across a fragmented spectrum, dedicated subsystems are developed in order to service individual frequency bands and specifications. Such a dedicated-hardware-intensive approach leads to high resource costs, including costs due to multiple instantiations of mixers, filters, and samplers. Such increases in hardware requirements in turn increases device size, power consumption, weight, and financial cost. If it can meet the required real-time constraints, a more flexible and reconfigurable design approach, such as a software-based solution, is often more desirable over a dedicated hardware solution. However, significant challenges must be overcome in order to meet constraints on throughput and latency while servicing different frequency bands and bandwidths. Graphics processing unit (GPU) technology provides a promising class of platforms for addressing these challenges. GPUs, which were originally designed for rendering images and video sequences, have been adapted as general purpose high-throughput computation engines for a wide variety of application areas beyond their original target domains. Linear algebra and signal processing acceleration are examples of such application areas. In this thesis, we apply GPUs as software-based, baseband radios and demonstrate novel, software-based implementations of key subsystems in modern wireless transceivers. In our work, we develop novel implementation techniques that allow communication system designers to use GPUs as accelerators for baseband processing functions, including real-time filtering and signal transformations. More specifically, we apply GPUs to accelerate several computationally-intensive, frontend radio subsystems, including filtering, signal mixing, sample rate conversion, and synchronization. These are critical subsystems that must operate in real-time to reliably receive waveforms. The contributions of this thesis can be broadly organized into 3 major areas: (1) channelization, (2) arbitrary resampling, and (3) synchronization. 1. Channelization: a wideband signal is shared between different users and channels, and a channelizer is used to separate the components of the shared signal in the different channels. A channelizer is often used as a pre-processing step in selecting a specific channel-of-interest. A typical channelization process involves signal conversion, resampling, and filtering to reject adjacent channels. We investigate GPU acceleration for a particularly efficient form of channelizer called a polyphase filterbank channelizer, and demonstrate a real-time implementation of our novel channelizer design. 2. Arbitrary resampling: following a channelization process, a signal is often resampled to at least twice the data rate in order to further condition the signal. Since different communication standards require different resampling ratios, it is desirable for a resampling subsystem to support a variety of different ratios. We investigate optimized, GPU-based methods for resampling using polyphase filter structures that are mapped efficiently into GPU hardware. We investigate these GPU implementation techniques in the context of interpolation (integer-factor increases in sampling rate), decimation (integer-factor decreases in sampling rate), and rational resampling. Finally, we demonstrate an efficient implementation of arbitrary resampling using GPUs. This implementation exploits specialized hardware units within the GPU to enable efficient and accurate resampling processes involving arbitrary changes in sample rate. 3. Synchronization: incoming signals in a wireless communications transceiver must be synchronized in order to recover the transmitted data properly from complex channel effects such as thermal noise, fading, and multipath propagation. We investigate timing recovery in GPUs to accelerate the most computationally intensive part of the synchronization process, and correctly align the incoming data symbols in the receiver. Furthermore, we implement fully-parallel timing error detection to accelerate maximum likelihood estimation.