检索结果-内蒙古大学图书馆

2023 International Conference on Electrical, Computer, Communications and Mechatronics Engineering, ICECCME 2023

作者： Van Den Brink, Arvid B. Meteer, Oǧuz Bekooij, Marco J.G. University of Twente Department of Computer Architectures for Embedded Systems Enschede Netherlands Nxp Semiconductors Department of Embedded Software and Signal Processing Eindhoven Netherlands

ISBN: (纸本)9798350322972

The belief propagation algorithm is one of the preferred algorithms for a polar code based decoder that can potentially achieve the lowest latency. Although belief propagation polar code algorithms have the ability for a highly parallelized implementation, they require more iterations to achieve a comparable frame error rate and bit error rate compared to the successive cancellation polar code algorithm. The iterative nature of the belief propagation algorithms also results in a higher computational complexity, i.e. O(N(2log2N-1)) compared to the computational complexity O(Nlog2N) of the successive cancellation decoder algorithm. In this paper we propose several software implementations of belief propagation polar code decoders using the computational complexity reduced belief propagation algorithm and the enhanced loop-weakened belief propagation algorithm, to increase the number of decoded codes per second. Compared to a baseline belief propagation implementation, our proposed radix-2 and radix-4 implementations increase the throughput by approximately 15.4% and 22.04% respectively. This gain is the result of a reduction of arithmetic operations, i.e., additions, compares, and multiplications, which is obtained without a loss in error-correcting performance. © 2023 IEEE.

关键词： Belief propagation

来源：评论

学校读者我要写书评

暂无评论

Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

Fusion Model Using a Neural Network and MLE for a Single Sna...

引用

2023 IEEE International Radar Conference, RADAR 2023

作者： De Oliveira, Marcio L. Lima Bekooij, Marco J. G. University of Twente Computer Architecture for Embedded Systems Enschede Netherlands University of Twente Department of Embedded Software and Signal Processing Nxp Semiconductors Eindhoven Netherlands

ISBN: (纸本)9781665482783

In this paper, we discuss the fusion of deep neural networks with the Maximum Likelihood Estimation (MLE) algorithm for estimating both the Direction Of Arrival (DOA) of MIMO radars and the number of sources for a single-snapshot scenario using a Minimum Redundancy antenna Array (MRA) with imperfections and evaluating it with synthetic and real-world data. The combination of Deep Learning (DL) with the classical MLE seems to be a viable solution for this problem, as it is less computationally intensive than a classical MLE while not losing generalization and being better at estimating the number of sources. In both our experiments, using synthetic and real-world data, our method performs close to MLE and appears to be a deployable solution for real scenarios. Besides reducing the computational load of MLE, the novelty of our model lies in the fact that the deep learning model learns the gain and phase errors as a function of the direction of arrival instead of applying a simple calibration. © 2023 IEEE.

关键词： Maximum likelihood estimation

来源：评论

学校读者我要写书评

暂无评论

Association of Camera and Radar Detections Using Neural Networks

Association of Camera and Radar Detections Using Neural Netw...

引用

IEEE National Conference on Radar

作者： Konstantinos Fatseas Marco J.G. Bekooij Computer Architectures for Embedded Systems University of Twente Enschede The Netherlands Embedded Software and Signal Processing NXP Semiconductors Eindhoven The Netherlands

Automotive radar and camera fusion relies on linear point transformations from one sensor's coordinate system to the other. However, these transformations cannot handle non-linear dynamics and are susceptible to sensor noise. Furthermore, they operate on a point-to-point basis, so it is impossible to capture all the characteristics of an object. This paper introduces a method that performs detection-to-detection association by projecting heterogeneous object features from the two sensors into a common high-dimensional space. We associate 2D bounding boxes and radar detections based on the Euclidean distance between their projections. Our method utilizes deep neural networks to transform feature vectors instead of single points. Therefore, we can leverage real-world data to learn non-linear dynamics and utilize several features to provide a better description for each object. We evaluate our association method against a traditional rule-based method, showing that it improves the accuracy of the association algorithm and it is more robust in complex scenarios with multiple objects.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Weakly Supervised Semantic Segmentation for Range-Doppler Maps 18

Weakly Supervised Semantic Segmentation for Range-Doppler Ma...

引用

18th European Radar Conference, EuRAD 2021

作者： Fatseas, Konstantinos Bekooij, Marco J.G. University of Twente Department of Computer Architectures for Embedded Systems Netherlands NXP Semiconductors Department of Embedded Software and Signal Processing Netherlands

ISBN: (纸本)9782874870651

Deep convolutional neural networks (DCNNs) have been successfully applied for object detection and semantic segmentation of radar range-Doppler (RD) maps. However, training a DCNN requires many annotated examples that are costly and difficult to create. In this work we present a method that reduces significantly the manual effort involved in the annotation of RD maps to train a DCNN for segmentation. A 40 times reduction in manual labelling effort is achieved because the annotation of each RD map includes only the class of the objects instead of drawing a polygon around the corresponding cells. The localization of the objects is performed by tracing back from the output to the input of a classification neural network. Experimental results show that our approach achieves robust localization performance in complex real-world urban scenarios as observed with a low-cost automotive radar. Furthermore, we show that our approach performs similarly to DCNNs that are trained with a publicly available dataset in which localization information is provided. © 2022 European Microwave Association (EuMA).

关键词： Convolutional neural networks

来源：评论

学校读者我要写书评

暂无评论

Software Implementation and Performance of a Computational Complexity Reduced Belief Propagation Polar Code Decoder

Software Implementation and Performance of a Computational C...

引用

Electrical, Computer, Communications and Mechatronics Engineering (ICECCME), International Conference on

作者： Arvid B. Van Den Brink Oğuz Meteer Marco J.G. Bekooij Department of Computer Architectures for Embedded Systems University of Twente Enschede The Netherlands Department of Embedded Software and Signal Processing NXP Semiconductors Eindhoven The Netherlands

The belief propagation algorithm is one of the preferred algorithms for a polar code based decoder that can potentially achieve the lowest latency. Although belief propagation polar code algorithms have the ability for a highly parallelized implementation, they require more iterations to achieve a comparable frame error rate and bit error rate compared to the successive cancellation polar code algorithm. The iterative nature of the belief propagation algorithms also results in a higher computational complexity, i.e. O(N(2log 2 N-1)) compared to the computational complexity O(Nlog 2 N) of the successive cancellation decoder algorithm. In this paper we propose several software implementations of belief propagation polar code decoders using the computational complexity reduced belief propagation algorithm and the enhanced loop-weakened belief propagation algorithm, to increase the number of decoded codes per second. Compared to a baseline belief propagation implementation, our proposed radix-2 and radix-4 implementations increase the throughput by approximately 15.4% and 22.04% respectively. This gain is the result of a reduction of arithmetic operations, i.e., additions, compares, and multiplications, which is obtained without a loss in error-correcting performance.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Fusion Model Using a Neural Network and MLE for a Single Snapshot DOA Estimation with Imperfection Mitigation

Fusion Model Using a Neural Network and MLE for a Single Sna...

引用

IEEE International Conference on Radar

作者： Marcio L. Lima De Oliveira Marco J. G. Bekooij Computer Architecture for Embedded Systems University of Twente Enschede The Netherlands Department of Embedded Software and Signal Processing NXP Semiconductors University of Twente Eindhoven The Netherlands

关键词：

来源：评论

学校读者我要写书评

暂无评论

A Panchromatic Image-based Spectral Imagery Super Resolution Algorithm

引用

Chinese Journal of Electronics 2023年第4期20卷 617-620页

作者： WANG Suyu ZHUO Li LI Xiaoguang Signal and Information Processing Laboratory Beijing University of Technology Beijing China Embedded Software and Systems Institute Beijing University of Technology Beijing China

A panchromatic image based spectral imagery super resolution algorithm is proposed to resolve the problem of remote sensing image resolution enhancement in this paper. A simple gray scale mapping method is first presented to better use high frequency details distilled from a high resolution reference panchromatic image, then a Maximum a posterior (MAP) based iterative optimization process is introduced to realize super resolution while keeping spectral characters. Evaluated by both simulated and really acquired images, the proposed method effectively improved the resolution of spectral imagery. PSNR (Peak signal to noise ratio) of the reconstructed images has improved l~5dB compared with the traditional bilinear interpolation and IHS (Intensity hue saturation) based method. Subjective quality of the enhanced images has obviously exceeded the other twos as well.

关键词： PSNR Superresolution Iterative methods Spatial resolution signal resolution Image reconstruction Image fusion

来源：评论

学校读者我要写书评

暂无评论

Challenges Using FPGA Clusters for Distributed CNN Training

Challenges Using FPGA Clusters for Distributed CNN Training

引用

International Conference on Field Programmable Logic and Applications

作者： Philipp Kreowsky Justin Knapheide Benno Stabernack Fraunhofer Institute for Telecommunications Heinrich Hertz Institute Berlin Germany Embedded Systems Architectures for Signal Processing University of Potsdam Potsdam Germany

While FPGAs are well-established in the field of CNN inference, there is a lack of research regarding CNN training. To mitigate this, we develop an easy-to-use framework for CNN training on network-attached FPGA clusters. Tightly pipelined layer parallelism promises to facilitate a high utilization across a whole cluster. This comes with challenges, which we discuss in this paper.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Demonstrating NADA: A Workflow for Distributed CNN Training on FPGA Clusters

Demonstrating NADA: A Workflow for Distributed CNN Training ...

引用

International Conference on Field Programmable Logic and Applications

作者： Justin Knapheide Philipp Kreowsky Benno Stabernack Fraunhofer Institute for Telecommunications Heinrich Hertz Institute Berlin Germany University of Potsdam Embedded Systems Architectures for Signal Processing Potsdam Germany

We introduce our Network Attached Deep learning Accelerator called NADA, which consists of a novel and flexible HW/SW framework for efficient training of deep neural networks on FPGA clusters. NADA is centered around layer parallelism, instantiating a specific implementation for each layer. These implementations are placed across the desired number of network attached FPGAs in the cluster. The NADA hardware architecture relies on a high-speed UDP/IP pure hardware network stack. We demonstrate the usability of our approach by training a couple of demo networks on an Arria 10 FPGA cluster.

关键词：

来源：评论

学校读者我要写书评

暂无评论

Energy-Efficient Radix-4 Belief Propagation Polar Code Decoding Using an Efficient Sign-Magnitude Adder and Clock Gating

Energy-Efficient Radix-4 Belief Propagation Polar Code Decod...

引用

Euromicro Symposium on Digital System Design

作者： Oğuz Meteer Arvid Van Den Brink Marco J.G. Bekooij Department of Computer Architectures for Embedded Systems University of Twente Enschede The Netherlands Department of Embedded Software and Signal Processing NXP Semiconductors Eindhoven The Netherlands

ISBN: (数字)9781665474047

ISBN: (纸本)9781665474054

Polar encoding is the first information coding method that has been proven to achieve channel capacity for binary-input discrete memoryless channels. Since its introduction, much research has been done on improving decoding performance, execution time and energy efficiency. Classic belief propagation uses radix-2 decoding, but a recent study proposed radix-4 decoding which reduces memory usage by 50%. However a drawback is its higher computational complexity, negatively impacting energy usage and throughput. In this paper we present an energy-efficient radix-4 belief propagation polar decoder architecture that uses a new sign-magnitude adder that does not require conversion to two's complement and back. On top of that we also propose using clock gating of input values by checking if all $R$ inputs of the decoder are zero. These two key contributions lead to a more energy -efficient design that is smaller and has higher maximum clock speed and throughput. Post-layout simulation results show that compared to the previously proposed 1024-bit radix-4 belief propagation polar code decoder, our decoder uses between 30.22 % and 32.80 % less power and is 5.2 % smaller at the same clock speed. Also, our design can achieve a 15.7% higher clock speed at which it is still up to 10.76% more power efficient and 4.8% smaller.

关键词： Simulation Switches Energy efficiency Encoding Decoding Polar codes Adders

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：