The proceedings contain 60 papers. The topics discussed include: extraction of fetal ECG from maternal abdominal record in the 3rd trimester of gestation using R-R interval windowing technique;energy modeling of clust...
ISBN:
(纸本)9781538663189
The proceedings contain 60 papers. The topics discussed include: extraction of fetal ECG from maternal abdominal record in the 3rd trimester of gestation using R-R interval windowing technique;energy modeling of cluster system;bandwidth efficient architectures for convolutional neural network;arithmetic computations based on chemical reaction networks;sparsity aware fast block LMS algorithms for MIMO radar imaging;design of iterative hybrid beamforming for multi-user mmWave massive MIMO systems;and efficient compressed landweber detector for massive MIMO.
Spiking neural networks have become an important family of neuron-based models that sidestep many of the key limitations facing modern-day backpropagation-trained deep networks, including their high energy inefficienc...
详细信息
ISBN:
(纸本)9798350373769;9798350373752
Spiking neural networks have become an important family of neuron-based models that sidestep many of the key limitations facing modern-day backpropagation-trained deep networks, including their high energy inefficiency and long-criticized biological implausibility. In this work, we design and investigate a proof-of-concept instantiation of contrastive-signal-dependent plasticity (CSDP), a neuromorphic form of forward-forward-based, backpropagation-free learning. Our experimental simulations demonstrate that a hardware implementation of CSDP is capable of learning simple logic functions without the need to resort to complex gradient calculations.
This paper presents a comprehensive study and implementations onto FPGA device of an Expectation Propagation (EP)-based receiver for QPSK, 8-PSK, and 16-QAM. To the best of our knowledge, this is the first for this ki...
详细信息
ISBN:
(纸本)9798350373769;9798350373752
This paper presents a comprehensive study and implementations onto FPGA device of an Expectation Propagation (EP)-based receiver for QPSK, 8-PSK, and 16-QAM. To the best of our knowledge, this is the first for this kind of receiver. The receiver implements a Frequency Domain (FD) Self-Iterated Linear Equalizer (SILE), where EP is used to approximate the true posterior distribution of the transmitted symbols with a simpler distribution. Analytical approximations for the EP feedback generation process and the three constellations are applied to lessen the complexity of the soft mapper/demapper architectures. The simulation results demonstrate that the fixed-point version performs comparably to the floating-point. Moreover, implementation results show the efficiency in terms of FPGA resource usage of the proposed architecture.
Sixth generation (6G) wireless networks envision new Internet-of-Things (IoT) applications that require faster and more secure data transmission. Reconfigurable intelligent surfaces (RISs) can improve physical-layer s...
详细信息
ISBN:
(纸本)9798350373769;9798350373752
Sixth generation (6G) wireless networks envision new Internet-of-Things (IoT) applications that require faster and more secure data transmission. Reconfigurable intelligent surfaces (RISs) can improve physical-layer security in wireless communications by optimally reflecting incoming electromagnetic waves to benefit legitimate users (LUs) and simultaneously hinder eavesdroppers (ED). However, for low-latency data transmission and short-coherence-time wireless channels, optimizing the RIS configuration must be done extremely fast, thus requiring a hardware accelerator. In this paper, we consider a multi-input multiple-output (MIMO) RIS-enabled physical-layer security system and introduce a novel low-latency hardware accelerator to maximize the achievable secrecy rate by optimizing the RIS reflection coefficients. The proposed accelerator implements a Dinkelbach-based iterative algorithm, here modified for the MIMO case and to facilitate hardware implementation. Simulations reveal that high-quality solutions are obtained, compared to a theoretical upper limit. The proposed sub-1-ms latency accelerator is detailed, highlighting design choices that enable its fast operation under moderate resource requirements, and also meeting the real-time requirements of emerging 6G applications.
The selection of an algorithm for MIMO detection often involves a trade-off between detection performance and the efficiency of the implementation. Tree-based algorithms, such as Sphere Decoding (SD), offer superior p...
详细信息
ISBN:
(纸本)9798350373769;9798350373752
The selection of an algorithm for MIMO detection often involves a trade-off between detection performance and the efficiency of the implementation. Tree-based algorithms, such as Sphere Decoding (SD), offer superior performance compared to linear schemes like MMSE and ZF and non-linear schemes like successive interference cancellation (SIC). In the SD approach, the algorithm must select the most suitable branch at each tree node. Due to this conditional behavior, the number of computational steps is data-dependent and is only determined at run-time. Moreover, due to the sequential nature of the tree traversal, it is not well suited for parallel implementation on a modern vector DSP. Approaches like K-Best exist to address this challenge, but they remain computationally intense, especially for higher-order QAM modulations. We study the use case of the 5G-based IoT standard NR-Redcap, which requires a low-complexity baseband kernel implementation for a cost-effective modem design. We propose a novel MIMO detection scheme called Box Decoding, whose computational complexity does not depend on the QAM modulation order and offers a significant reduction in complexity compared to K-Best for our chosen use case. Furthermore, the algorithm doesn't necessitate the expensive sorting of candidate symbols, making it particularly appealing for implementation on contemporary vector Digital signal Processors (DSPs). Testing on a 64QAM modulation system for Redcap indicates negligible MIMO detection performance degradation compared to K-Best, with significantly lower complexity. For our use case, the Box Decoding algorithm proves to be a cost-effective implementation scheme for MIMO detection in 5G IoT modems, offering a 2dB SNR gain over MMSE-based SIC for an extra 23% of clock cycles consumed for this task on a 512-bit vector DSP.
The proceedings contain 36 papers. The topics discussed include: evaluation of coarse-grained reconfigurable array for a dual mode OTFS-OFDM modulator;fast energy optimization of on-chip ECC memories;box decoding: a l...
ISBN:
(纸本)9798350373752
The proceedings contain 36 papers. The topics discussed include: evaluation of coarse-grained reconfigurable array for a dual mode OTFS-OFDM modulator;fast energy optimization of on-chip ECC memories;box decoding: a low-complexity algorithm for MIMO detection;HM-detect: a novel method for heart murmur detection and classification using machine learning and sound feature analysis;on the mismatch between the phase structure of all-pole-based synthetic vowels and natural vowels;QEEGNet: quantum machine learning for enhanced electroencephalography encoding;processing multi-layer perceptrons in-memory;hierarchical quantum control gates for functional MRI understanding;and noise identification for data-augmented physics-based state-space models.
The proceedings contain 58 papers. The topics discussed include: budget restricted incremental learning with pre-trained convolutional neural networks and binary associative memories;a modified gradient descent bit fl...
ISBN:
(纸本)9781538604465
The proceedings contain 58 papers. The topics discussed include: budget restricted incremental learning with pre-trained convolutional neural networks and binary associative memories;a modified gradient descent bit flipping decoding scheme for LDPC codes;a stochastic number representation for fully homomorphic cryptography;a reduction of circuit size of digital direct-driven speaker architecture using segmented pulse shaping technique;CRN-based design methodology for synchronous sequential logic;an efficient conjugate residual detector for massive MIMO systems;analysing the performance of divide-and-conquer sequential matrix diagonalisation for large broadband sensor arrays;blind detection of polar codes;a discriminative spectral-temporal feature set for motor imagery classification;advanced wireless digital baseband signalprocessing beyond 100 Gbit/S;a split pre-conditioned conjugate gradient method for massive MIMO detection;a joint spatial texture analysis/watermarking system for digital image authentication;digital self-interference cancellation in inter-band carrier aggregation transceivers: algorithm and digital implementation perspectives;design space exploration of dataflow-based Smith-Waterman FPGA implementations;and DVFS based power management for LDPC decoders with early termination.
The proceedings contain 32 papers. The topics discussed include: Bayesian optimization for simultaneous deconvolution of room impulse responses;beam alignment for the cell-free mmWave massive MU-MIMO uplink;ultra-fast...
ISBN:
(纸本)9781665485241
The proceedings contain 32 papers. The topics discussed include: Bayesian optimization for simultaneous deconvolution of room impulse responses;beam alignment for the cell-free mmWave massive MU-MIMO uplink;ultra-fast machine learning inference through c code generation for tangled program graphs;cosmic: a real-time platform for signalprocessing pipelines;common spatial pattern with deep learning for fetal heart rate monitoring;virtual triggering: a technique to segment cryptographic processes in side-channel traces;statistical and morphological component separation of foregrounds in convolved HI skymaps;multi-factor pruning for recursive projection-aggregation decoding of RM codes;and an adjustable farthest point sampling method for approximately-sorted point cloud data.
Sensors are used to monitor various parameters in many real-world applications. Sudden changes in the underlying patterns of the sensors readings may represent events of interest. Therefore, event detection, an import...
详细信息
ISBN:
(纸本)9781665401449
Sensors are used to monitor various parameters in many real-world applications. Sudden changes in the underlying patterns of the sensors readings may represent events of interest. Therefore, event detection, an important temporal version of outlier detection, is one of the primary motivating applications in sensor networks. This work describes the implementation of a real-time outlier detection that uses an Autoencoder-LSTM neural-network accelerator implemented on the Xilinx PYNQ-Z1 development board. The implemented accelerator consists of a fine-tuned Autoencoder to extract the latent features in sensor data followed by a Long short-term memory (LSTM) network to predict the next step and detect outliers in real-time. The implemented design achieves 2.06 ms minimum latency and 85.9 GOp/s maximum throughput. The low latency and 0.25 W power consumption of the Autoencoder-LSTM outlier detector makes it suitable for resource-constrained computing platforms.
暂无评论