http://hdl.handle.net/1903/2234 2013-05-21T12:17:59Z http://hdl.handle.net/1903/13890 Title: Empirical Speedup Study of Truly Parallel Data Compression Authors: Edwards, James A.; Vishkin, Uzi Abstract: We present an empirical study of novel work-optimal parallel algorithms for Burrows-Wheeler compression and decompression of strings over a constant alphabet. To validate these theoretical algorithms, we implement them on the experimental XMT computing platform developed especially for supporting parallel algorithms at the University of Maryland. We show speedups of up to 25x for compression, and 13x for decompression, versus bzip2, the de facto standard implementation of Burrows-Wheeler compression. Unlike existing approaches, which assign an entire (e.g., 900KB) block to a processor that processes the block serially, our approach is “truly parallel” as it processes in parallel the entire input. Besides the theoretical interest in solving the “right” problem, the importance of data compression speed for small inputs even at great expense of quality (compressed size of data) is demonstrated by the introduction of Google’s Snappy for MapReduce. Perhaps surprisingly, we show feasibility of holding on to quality, while even beating Snappy on speed. In turn, this work adds new evidence in support of the XMT/PRAM thesis: that an XMT-like many-core hardware/ software platform may be necessary for enabling general-purpose parallel computing. Comparison of our results to recently published work suggests 70x improvement over what current commercial parallel hardware can achieve. 2013-04-20T00:00:00Z http://hdl.handle.net/1903/13872 Title: Design and testing methodologies for signal processing systems using DICE Authors: Kedilaya, Soujanya Akirebari Abstract: The design and integration of embedded systems in heterogeneous programming environments is still largely done in an ad hoc fashion making the overall development process more complicated, tedious and error-prone. In this work, we propose enhancements to existing design flows that utilize model-based design to verify cross-platform correctness of individual actors. The DSPCAD Integrative Command Line Environment (DICE) is a realization of managing these enhancements. We demonstrate this design flow with two case studies. By using DICE's novel test framework on modules of a triggering system in the Large Hadron Collider, we demonstrate how the cross-platform model-based approach, automatic testbench creation and integration of testing in the design process alleviate the rigors of developing such a complex digital system. The second case study is an exploration study into the required precision for eigenvalue decomposition using the Jacobi algorithm. This case study is a demonstration of the use of dataflow modeling in early stage application exploration and the use of DICE in the overall design flow. 2010-01-01T00:00:00Z http://hdl.handle.net/1903/13861 Title: Decoding of walking kinematics from non-invasively acquired electroencephalographic signals in stroke patients Authors: Nathan, Kevin Abstract: Our group has recently shown the feasibility of decoding kinematics of controlled walking from the lower frequency range of electroencephalographic (EEG) signals during a precision walking task. Here, we turn our attention to stroke survivors who have had lesions resulting in hemiparetic gait. We recorded the EEG of stroke recovery patients during a precision treadmill walking task while tracking bilaterally the kinematics of the hips, knees, and ankles. In offline analyses, we applied a Wiener Filter and two unscented Kalman filters of 1st and 10th orders to predict estimates of the kinematic parameters from scalp EEG. Decoding accuracies from four patients who have had cortical and subcortical strokes were comparable with previous studies in healthy subjects. With improved decoding of EEG signals from damaged brains, we hope we can soon correlate activity to more intentional and normal-form walking that can guide users of a powered lower-body prosthetic or exoskeleton. 2012-01-01T00:00:00Z http://hdl.handle.net/1903/13854 Title: AN ATOMIC LAYER DEPOSITION PASSIVATED SURFACE ACOUSTIC WAVE SENSOR FOR BACTERIAL BIOFILM GROWTH MONITORING Authors: Kim, Young Wook Abstract: This thesis reports for the first time the design, fabrication, and testing of a reusable Surface Acoustic Wave (SAW) sensor for biofilm growth monitoring. Bacterial biofilms cause severe infections, and are often difficult to remove without an invasive surgery. Thus, their detection at an early stage is critical for effective treatments. A highly sensitive SAW sensor for biofilm growth monitoring was fabricated by depositing a high quality zinc oxide (ZnO) piezoelectric thin film by pulsed laser deposition (PLD). The sensor was successfully passivated by aluminum oxide (Al2O3) using Atomic Layer Deposition (ALD) to prevent ZnO damage from long term media contact. The sensor was reusable over multiple biofilm formation experiments using the ALD Al2O3 passivation and an oxygen plasma biofilm cleaning method. The SAW sensor was studied with Escherichia coli biofilm growth in Lysogeny Broth (LB) and in 10% Fetal Bovine Serum (FBS) as a simulated an in vivo environment. A multiple MHz level resonant frequency shift measured at the output of the SAW sensor in both LB and 10% FBS corresponded to the natural biofilm growth progression. These repeatable E. coli biofilm growth monitoring results validate the novel application of a SAW sensor for future implantable biofilm sensing applications. 2012-01-01T00:00:00Z