Search

The repeated program/erase (P/E) cycles wear out flash memory cells which limits the device’s lifetime. The larger the injected/erased charge, the faster the wear out rate and thus there is interest in reducing the percentage of cells programmed to high voltage levels. Conventional data shaping techniques often introduce a nonnegligible overhead as well as incurring additional complexity to encode the input data into shaped codewords and decode the read data. We propose a novel lossless yet simple endurance modulation technique that relies solely on the error-correction code (ECC) to retrieve the original data. This idea is motivated by the fact that ECCs are usually designed to correct channel errors towards the end of the device’s lifetime and hence at early age, the ECC can be jointly used for channel error correction and endurance demodulation. The proposed scheme consists of data shaping by flipping bits in a controlled manner and simple max-log demodulation to recover those flipped bits. The simulation results show that the proposed scheme does not undermine the device’s ability to recover the original data if the shaping gain is adjusted as the device ages. Once the channel errors exceed a prescribed threshold, the controller disables the endurance modulation and operates in normal mode. [ABSTRACT FROM AUTHOR]

As the density of flash memory is increasing to improve storage capacity, it also becomes more vulnerable to reading errors. Hence it is common practice to employ error-correcting codes, such as low-density parity check (LDPC) codes. To exploit the error-correcting capability of these codes, soft input must be provided to the decoder. This means that the flash memory controller needs to be able to provide reliability values, which is usually obtained through multiple ‘soft’ reads of the cells in the form of log-likelihood ratios (LLR). However finding these reliabilities with high accuracy is not easy with very few reads. In this brief we present a novel approach to estimate the LLRs based on a previous idea by Sharon et al. for regular LDPC codes. We present a low-complexity solution using machine learning and also an extension to irregular LDPC codes. A neural network is trained with offline data (simulated or measured) for a wide range of operating conditions and can then be applied to flash memory readings. By simulations we show that this provides good performance with only a small loss compared with genie-aided LLRs. [ABSTRACT FROM AUTHOR]

The incorporation of powerful soft-decision error correction codes (e.g., LDPC) to replace legacy hard-decision codes (e.g., BCH) can boost the reliability of flash memory beyond its current limits at the expense of additional reading latency to generate soft information, namely in the form of log-likelihood ratios (LLRs). Unfortunately, the memory channel varies significantly depending on several factors (e.g., temperature, parasitic coupling, and retention time) which renders the use of precomputed LLR lookup table (LUT) ineffective especially as the device ages. In fact, the estimated LLRs need to be regularly updated to avoid any performance degradation. We provide a dynamic yet low-latency blind detection algorithm which is able to adaptively adjust the hard/soft-decision boundaries and estimate the corresponding LLR codebook. The proposed demodulator offers near-optimal performance with low-complexity overhead. [ABSTRACT FROM AUTHOR]

To provide reliability of flash memory, error-correcting codes are usually used which can be further enhanced by moving on from hard-decision to soft-decision decoders. This requires multiple readings to obtain a reliability measure of the stored data, which is complicated by the fact that a posteriori probability distributions change as the memory ages through retention and wear-out. The existing methods for reading a flash memory either assume knowledge of these distributions or require a large number of readings for accurate estimation. In this letter, we propose a low-complexity algorithm that tracks the soft-decision boundaries and estimates the corresponding log-likelihood ratios (LLRs). It does not require any additional readings and only involves simple calculations using already available variables. The efficacy is verified through the simulations and is shown to provide near-optimal performance. [ABSTRACT FROM AUTHOR]

Write-once memory (WOM) is a storage device, where the cells can only be increased when programmed. A WOM code allows the device to be programmed multiple times without decreasing the cell values. This will allow a longer lifetime in, e.g., flash memory, where frequent erasures wear it out. Most WOM codes assume that the previous state is available when encoding a message; however, this is impractical as it requires an additional read operation before encoding can be performed. In this letter, we will consider the case where neither the encoder nor the decoder knows the previous state of the memory. Furthermore, we will present a WOM code for nonbinary cells, i.e., where each cell can store $q$ different values. We will derive the sum rate of this code and show that it exceeds that of the previous designs. [ABSTRACT FROM PUBLISHER]

Complex Orthogonal Design (COD) codes are known to have the lowest detection complexity among Space-Time Block Codes (STBCs). However, the rate of square COD codes decreases exponentially with the number of transmit antennas. The Quasi-Orthogonal Design (QOD) codes emerged to provide a compromise between rate and complexity as they offer higher rates compared to COD codes at the expense of an increase of decoding complexity through partially relaxing the orthogonality conditions. The QOD codes were then generalized with the so called g-symbol and g-group decodable STBCs where the number of orthogonal groups of symbols is no longer restricted to two as in the QOD case. However, the adopted approach for the construction of such codes is based on sufficient but not necessary conditions which may limit the achievable rates for any number of orthogonal groups. In this paper, we limit ourselves to the case of Unitary Weight (UW)-g-group decodable STBCs for 2^a transmit antennas where the weight matrices are required to be single thread matrices with non-zero entries ∈{± 1,± j} and address the problem of finding the highest achievable rate for any number of orthogonal groups. This special type of weight matrices guarantees full symbol-wise diversity and subsumes a wide range of existing codes in the literature. We show that in this case an exhaustive search can be applied to find the maximum achievable rates for UW-g-group decodable STBCs with g>1. For this purpose, we extend our previously proposed approach for constructing UW-2-group decodable STBCs based on necessary and sufficient conditions to the case of UW-g-group decodable STBCs in a recursive manner. [ABSTRACT FROM AUTHOR]

Space-time coding techniques have become common-place in wireless communication standards as they provide an effective way to mitigate the fading phenomena inherent in wireless channels. However, the use of Space-Time Block Codes (STBCs) increases significantly the optimal detection complexity at the receiver unless the low complexity decodability property is taken into consideration in the STBC design. In this letter we propose a new low-complexity decodable rate-1 full-diversity 4x 4 STBC. We provide an analytical proof that the proposed code has the Non-Vanishing-Determinant (NVD) property, a property that can be exploited through the use of adaptive modulation which changes the transmission rate according to the wireless channel quality. We compare the proposed code to existing low-complexity decodable rate-1 full-diversity 4x 4 STBCs in terms of performance over quasi-static Rayleigh fading channels, detection complexity and Peak-to-Average Power Ratio (PAPR). Our code is found to provide the best performance and the smallest PAPR which is that of the used QAM constellation at the expense of a slight increase in detection complexity w.r.t. certain previous codes but this will only penalize the proposed code for high-order QAM constellations. [ABSTRACT FROM PUBLISHER]