FPGA Architecture for 2D Discrete Fourier Transform Based on 2D Decomposition for Large-sized Data

Applications based on Discrete Fourier Transforms (DFT) are extensively used in several areas of signal and digital image processing. Of particular interest is the two-dimensional (2D) DFT which is more computation- and bandwidth-intensive than the one-dimensional (1D) DFT. Traditionally, a 2D DFT is computed using Row-Column (RC) decomposition, where 1D DFTs are computed along the rows followed by 1D DFTs along the columns. Both application specific and reconfigurable hardware have utilized this scheme for high-performance implementations of 2D DFT. However, architectures based on RC decomposition are not efficient for large input size data due to memory bandwidth constraints. In this paper, we propose an efficient architecture to implement 2D DFT for large-sized input data based on a novel 2D decomposition algorithm. This architecture achieves very high throughput by exploiting the inherent parallelism due to the algorithm decomposition and by utilizing the row-wise burst access pattern of the external memory. A high throughput memory interface has been designed to enable maximum utilization of the memory bandwidth. In addition, an automatic system generator is provided for mapping this architecture onto a reconfigurable platform of Xilinx Virtex-5 devices. For a 2K ×2K input size, the proposed architecture is 1.96 times faster than RC decomposition based implementation under the same memory constraints, and also outperforms other existing implementations.     

Algorithm and Parallel Implementation of Particle Filtering and its Use in Waveform-Agile Sensing

Sequential Monte Carlo particle filters (PFs) are useful for estimating nonlinear non-Gaussian dynamic system parameters. As these algorithms are recursive, their real-time implementation can be computationally complex. In this paper, we analyze the bottlenecks in existing parallel PF algorithms, and propose a new approach that integrates parallel PFs with independent Metropolis–Hastings (PPF-IMH) resampling algorithms to improve root mean-squared estimation error (RMSE) performance. We implement the new PPF-IMH algorithm on a Xilinx Virtex-5 field programmable gate array (FPGA) platform. For a one-dimensional problem with 1,000 particles, the PPF-IMH architecture with four processing elements uses less than 5% of a Virtex-5 FPGA’s resource and takes 5.85 μs for one iteration. We also incorporate waveform-agile tracking techniques into the PPF-IMH algorithm. We demonstrate a significant performance improvement when the waveform is adaptively designed at each time step with 6.84 μs FPGA processing time per iteration.     

Improving the Reliability of MLC NAND Flash Memories Through Adaptive Data Refresh and Error Control Coding

NAND Flash memory has become the most widely used non-volatile memory technology. We focus on multi-level cell (MLC) NAND Flash memories because they have high storage density. Unfortunately MLC NAND Flash memory also has reliability problems due to narrower threshold voltage gap between logical states. Errors in these memories can be classified into data retention (DR) errors and program interference (PI) errors. DR errors are dominant if the data storage time is longer than 1 day and these errors can be reduced by refreshing the data. PI errors are dominant if the data storage time is less than 1 day and these errors can be handled by error control coding (ECC). In this paper we propose a combination of data refresh policies and low cost ECC schemes that are cognizant of application characteristics to address the errors in MLC NAND Flash memories. First, we use Gray code based encoding to reduce the error rates in the four subpages (MSB-even, LSB-even, MSB-odd, LSB-odd) of a 2-bit MLC NAND Flash memory. Next, we apply data refresh techniques where the refresh interval is a function of the program/erase (P/E) frequency of the application. We show that an appropriate choice of refresh interval and BCH based ECC scheme can minimize memory energy while satisfying the reliability constraint.     

Apple quality,storage, and washing treatments affect patulin levels in apple cider

Patulin is a mycotoxin produced primarily by Penicillium expansum, a mold responsible for rot in apples and other fruits. The growth of this fungus and the production of patulin are common in fruit that has been damaged. However, patulin can be detected in visibly sound fruit. The purpose of this project was to determine how apple quality, storage, and washing treatments affect patulin levels in apple cider. Patulin was not detected in cider pressed from fresh tree-picked apples (seven cultivars) but was found at levels of 40.2 to 374 microg/liter in cider pressed from four cultivars of fresh ground-harvested (dropped) apples. Patulin was not detected in cider pressed from culled tree-picked apples stored for 4 to 6 weeks at 0 to 2 degrees C but was found at levels of 0.97 to 64.0 microg/liter in cider pressed from unculled fruit stored under the same conditions. Cider from controlled-atmosphere-stored apples that were culled before pressing contained 0 to 15.1 microg of patulin per liter, while cider made from unculled fruit contained 59.9 to 120.5 microg of patulin per liter. The washing of ground-harvested apples before pressing reduced patulin levels in cider by 10 to 100%, depending on the initial patulin levels and the type of wash solution used. These results indicate that patulin is a good indicator of the quality of the apples used to manufacture cider. The avoidance of ground-harvested apples and the careful culling of apples before pressing are good methods for reducing patulin levels in cider.     

Hardware Acceleration for Neuromorphic Vision Algorithms

Neuromorphic vision algorithms are biologically inspired models that follow the processing that takes place in the primate visual cortex. Despite their efficiency and robustness, the complexity of these algorithms results in reduced performance when executed on general purpose processors. This paper proposes an application-specific system for accelerating a neuromorphic vision system for object recognition. The system is based on HMAX, a biologically-inspired model of the visual cortex. The neuromorphic accelerators are validated on a multi-FPGA system. Results show that the neuromorphic accelerators are 13.8× (2.6×) more power efficient when compared to CPU (GPU) implementation.     

Binding free energy calculations of galectin-3-ligand interactions

Galectins show remarkable binding specificity towards ß-galactosides.A recently developed method for calculating binding free energiesbetween a protein and its substrates has been used to evaluatethe binding specificity of galectin-3. Five disaccharides anda tetrasaccharide were used as the substrates. The calculatedbinding free energies agree quite well with the experimentaldata and the ranking of binding affinities is well reproduced.For all the six protein–ligand complexes it was observedthat electrostatic interactions oppose binding whereas the non-polarcontributions drive complex formation. The observed bindingspecificity of galectin-3 for galactosides rather than glucosidesis discussed in light of our results.     

Estimation of relative evapotranspiration from NOAA PAL to derive growth characteristics in India

Clear‐sky dekadal relative evapotranspiration (RET) was derived using the surface energy‐balance approach applied to 10‐day composite NOAA PAL (8 km×8 km) datasets over the Indian landmass. This was further used to differentiate between growth characteristics for an irrigated intensive agriculture over a northern India state (e.g. Punjab) and a rainfed ill‐posed agriculture over a central India state (e.g. Madhya Pradesh) using time‐series data sets for five growing years (June–April): 1996–1997, 1997–1998, 1998–1999, 1999–2000, and 2000–2001. The triangular scatter between RET and normalized difference vegetation index (NDVI) showed that the minimum RET increases linearly with NDVI producing a ‘basal line’ that represents relative canopy transpiration only. A clear distinction in scatter was found between the two contrasting agro‐ecosystems showing a higher RET or root zone wetness in irrigated than rainfed systems. In rainfed rice‐growing regions, an inverse correlation (0.6–0.75) was found between RET and the Keetch–Byram meteorological drought index (KBDI), and a substantial reduction in RET was also found in a sub‐normal (2000) compared with a normal (1999) monsoon season. RET estimates were found to be most sensitive to atmospheric transmissivity followed by other land‐surface radiation budget inputs, such as NDVI, LST, and albedo. Error propagation due to three surface parameters is the opposite of that for transmissivity. The maximum possible error in clear‐sky NOAA PAL RET was estimated to be 12–15%. This test study would be helpful in deriving RET using optical and thermal data from a suite of current and future Indian geostationary satellite sensors for monitoring growing conditions.     

Parallel High Throughput Soft-Output Sphere Decoding Algorithm

Multiple-Input-Multiple-Output communication systems demand fast sphere decoding with high performance. To speed up the computation, we propose a scheme with multiple fixed complexity sphere decoders to construct a parallel soft-output fixed complexity sphere decoder (PFSD). The proposed decoder is highly parallel and has performance comparable to soft-output list fixed complexity sphere decoder (LFSD) and K-best sphere decoder. In addition, we propose a parallel QR decomposition algorithm to lower the preprocessing overhead, and a low complexity LLR algorithm to allow parallel update of LLR values. We demonstrate that the PFSD algorithm can increase the throughput and reduce bit error rate of a soft-output solution in a 4 × 4 16-QAM system, and has superior performance compared to other soft decoders with comparable throughput and computation complexity. The PFSD algorithm has been mapped onto Xilinx XC4VLX160 FPGA. The resulting PFSD decoder can achieve up to 75 Mbps throughput for 4 × 4 64-QAM configuration at 100MHz with low control overhead.