检索结果-内蒙古大学图书馆

A Numerical Approximation Approach for Deriving computational efficient Inverse Dynamics of 6-DOF Parallel Robots Based on Principle of Virtual Work

引用

IEEE ROBOTICS AND AUTOMATION LETTERS 2024年第6期9卷 5966-5973页

作者： Liu, Zhihua Cai, Chenguang Yang, Ming Shao, Zhufeng Lv, Qi Natl Inst Metrol Beijing 100029 Peoples R China Guizhou Univ Elect Engn Coll Guiyang 550025 Peoples R China Tsinghua Univ Dept Mech Engn State Key Lab Tribol Beijing 100086 Peoples R China

The inverse dynamics of the six degree-of-freedom (6-DOF) parallel robot (PR) presents an inherent complexity due to the closed-loop kinematic chains. To derive computational efficient inverse dynamics for real-time control, this study presents a numerical approximation (NA) approach based on the principle of virtual work. The analytical integration of the inertial wrenches is conducted for more concise formulas and then the numerical differentiation of these formulas is presented for more computational efficient inertial wrenches. The principle of virtual work is applied to derive the inverse dynamics by using the numerical inertial wrenches together with the analytical Jacobian matrices. The NA approach can reduce the computational cost of the inverse dynamics by up to 27% with a considerable computational accuracy. The real-time experimental results demonstrate that the NA approach has shortened the time interval of driving force compensation and significantly reduced the trajectory tracking error from 2.59 mm to 0.51 mm.

关键词： computational efficient inverse dynamics numerical approximation parallel robot

来源：评论

学校读者我要写书评

暂无评论

computational efficient segmented integration method for high-speed maneuvering target detection

引用

SIGNAL PROCESSING 2022年 195卷 108475-108475页

作者： Chen, Haixu Li, Xiaolong Wang, Kaiyao Sun, Zhi Cui, Guolong Univ Elect Sci & Technol China Sch Informat & Commun Engn Chengdu 611731 Sichuan Peoples R China

Long-time integration can improve the radar detection performance of the moving targets. However, the high speed maneuvering targets (HSMT) would cause range migration (RM) and Doppler Spread (DS) during the long integration process, which would lead to the failure of traditional moving target detection (MTD) method. The existed Generalized Radon-Fourier transform (GRFT) method can achieve a good integration and detection performance using multi-dimensional searching among the target parameter space, but the computational burden is too large. We propose a computational efficient segmented integration method for the HSMT detection, which could reduce the computational cost with slight integration performance loss. First, the echo signal is segmented based on the criterion that the RM and DS caused by the acceleration in each sub-segment can be ignored. After that, the proposed method obtains the integration of each sub-segment through Frequency bin Randon Fourier transform (FBRFT), while FBRFT can be quickly realized through Chirp-z transform (CZT). Subsequently, by analyzing the amplitude characteristics and phase characteristics of the FBRFT output, a matched accumulation method is presented to coherently integrate the energy among different sub-segments to further improve the signal-to-noise ratio (SNR). Finally, the performance of the proposed algorithm is analyzed through simulation experiments.(c) 2022 Elsevier B.V. All rights reserved.

关键词： Long time integration Segmented integration computational efficient Maneuvering target

来源：评论

学校读者我要写书评

暂无评论

Improved joint modeling of longitudinal and survival data using a poisson regression approach

引用

STATISTICAL METHODS AND APPLICATIONS 2025年 1-20页

作者： Liu, Yixiu Torabi, Mahmoud Zhang, Xuekui Jiang, Depeng Univ Manitoba Dept Community Hlth Sci 753 McDermot Ave Winnipeg MB R3E 0T6 Canada Univ Manitoba Dept Stat 186 Dysart Rd Winnipeg MB R3T 2N2 Canada Univ Victoria Dept Math & Stat 3800 Finnerty Rd Victoria BC V8P 5C2 Canada

Data of repeated measurements (longitudinal) and time-to-events (survival) are commonly recorded in studies. The joint model (JM) of longitudinal and survival data, which allows simultaneously analysis of the two types of outcomes, has been extensively discussed recently. JMs are computationally intensive due to large number of parameters and the complexity of fitting the survival submodel. The centerpiece of the survival submodel is the piecewise constant proportional hazard (PCPH). An alternative to PCPH for analysing survival data is the auxiliary Poisson regression model. However, the use of this approach in JMs has not been discussed. In this study, we propose using the auxiliary Poisson model as the survival part in a JM within a Bayesian framework. We conducted comprehensive simulation studies to assess the performance of our proposed method under various conditions and compared it to a published R package for JMs called JMbayes. Additionally, we used data from the Manitoba Follow-Up Study to illustrate the advantages and feasibility of our proposed method. The findings have showed that using the auxiliary Poisson approach as the survival submodel is a very promising method for jointly modeling longitudinal and survival data, as it helps decrease the computing burden.

关键词： Joint model of longitudinal and survival data Poisson regression model Bayesian estimation computational efficient Cox regression model

来源：评论

学校读者我要写书评

暂无评论

COMPUSR: computationally efficient Unsupervised Super-Resolution Approach for Wireless Capsule Endoscopy 8th

COMPUSR: Computationally Efficient Unsupervised Super-Resolu...

引用

8th International Conference on Computer Vision and Image Processing (CVIP)

作者： Saha, Nirban Sarvaiya, Anjali Upla, Kishor Raja, Kiran Pedersen, Marius Sardar Vallabhbhai Natl Inst Technol SVNIT Surat India Norwegian Univ Sci & Technol NTNU Gjovik Norway

ISBN: (纸本)9783031581731;9783031581748

Wireless Capsule Endoscopy (WCE) is an imaging technology for diseases related to Gastrointestinal (GI) track. However, due to limited hardware, the spatial resolution of acquired images is usually coarser leading to poor diagnostic quality. To enhance high-frequency details associated to edges, object boundaries, corners, etc., which are required for detection and recognition tasks, High-Resolution (HR) images are usually constructed through software-driven Super-Resolution (SR) techniques. Deep learning-based SR approaches has been frequently explored in the computer-vision community recently to increase the SR picture's quality. However, the prime barrier associated with the presently available deep learning SR approaches is the usage of supervised training where, LR images are simulated by using known deterioration (for instance, bicubic downsampling) from the acquired HR images. It is apparent that the deep models trained on such LR-HR pair exhibits poor generalization on the real LR observation. On the other hand, it is challenging to collect large clinical datasets consisting of true LR-HR pair for supervised learning. To circumvent this problem, we propose a computationally efficient unsupervised approach for SR task using Generative Adversarial Network (GAN) for WCE images (referred as COMPUSR). The proposed model avoids LR-HR pair and is trained using unsupervised strategy to learn the degradation of LR observation. It is validated on a derivative dataset of the original Kvasir dataset consisting of 10, 000 training samples and reveals considerable improvement over the other state-of-the-art methods on subjective and quantitative evaluations. Remarkably, the proposed approach needs significantly less number of learnable parameters (i.e., 2.4M) and a number of GFLOPS (i.e., 91.56G) than that of other approaches.

关键词： Super-Resolution Wireless Capsule Endoscopy Unsupervised Approach GAN computational efficient

来源：评论

学校读者我要写书评

暂无评论

Faster and Stronger Lossless Compression with Optimized Autoregressive Framework 60

Faster and Stronger Lossless Compression with Optimized Auto...

引用

60th ACM/IEEE Design Automation Conference (DAC)

作者： Mao, Yu Li, Jingzong Cui, Yufei Xue, Jason Chun City Univ Hong Kong Dept Comp Sci Hong Kong Peoples R China McGill Univ Sch Comp Sci Montreal PQ Canada

ISBN: (纸本)9798350323481

Neural AutoRegressive (AR) framework has been applied in general-purpose lossless compression recently to improve compression performance. However, this paper found that directly applying the original AR framework causes the duplicated processing problem and the in-batch distribution variation problem, which leads to deteriorated compression performance. The key to address the duplicated processing problem is to disentangle the processing of the history symbol set at the input side. Two new types of neural blocks are first proposed. An individual-block performs separate feature extraction on each history symbol while a mix-block models the correlation between extracted features and estimates the probability. A progressive AR-based compression framework (PAC) is then proposed, which only requires one history symbol from the host at a time rather than the whole history symbol set. In addition, we introduced a trainable matrix multiplication to model the ordered importance, replacing previous hardware-unfriendly Gumble-Softmax sampling. The in-batch distribution variation problem is caused by AR-based compression's structured batch construction. Based on this observation, a batch-location-aware individual block is proposed to capture the heterogeneous in-batch distributions precisely, improving the performance without efficiency losses. Experimental results show the proposed framework can achieve an average of 130% speed improvement with an average of 3% compression ratio gain across data domains compared to the state-of-the-art.

关键词： auto-regressive general-purpose lossless data compression hardware friendly neural networks computational efficient

来源：评论

学校读者我要写书评

暂无评论

Robust, Effective and Resource efficient Deep Neural Network for Intrusion Detection in IoT Networks 8

Robust, Effective and Resource Efficient Deep Neural Network...

引用

8th ACM Cyber-Physical System Security Workshop

作者： Zakariyya, Idris Kalutarage, Harsha Al-Kadri, M. Omar Robert Gordon Univ Sch Comp Aberdeen Scotland Birmingham City Univ Sch Comp & Digital Technol Birmingham W Midlands England

ISBN: (纸本)9781450391764

Internet of Things (IoT) devices are becoming increasingly popular and an integral part of our everyday lives, making them a lucrative target for attackers. These devices require suitable security mechanisms that enable robust and effective detection of attacks. Deep Neural Networks (DNNs) offer a promise, but they require large amounts of computational resources to provide better detection, and their detection capabilities can be exploited by adversarial attacks. Therefore, this paper proposes a method to train Fully Connected Neural Network (FCNN) for IoT security monitoring in a robust, effective and resource-efficient way. The resulting model is assessed against various benchmark datasets created using commercial IoT devices, such as doorbells, security cameras, and thermostats. Experimental results demonstrate the model's ability to maintain state-of-the-art accuracy and F1-score while reducing training memory and time consumption by 99.99 and 99.80 percentage points than its benchmark counterpart.

关键词： IoT Intrusion detection Deep neural network computational efficient IDS

来源：评论

学校读者我要写书评

暂无评论

Accelerating General-Purpose Lossless Compression via Simple and Scalable Parameterization 22

Accelerating General-Purpose Lossless Compression via Simple...

引用

30th ACM International Conference on Multimedia (MM)

作者： Mao, Yu Cui, Yufei Kuo, Tei-Wei Xue, Chun Jason City Univ Hong Kong Dept Comp Sci Hong Kong Peoples R China McGill Univ Sch Comp Sci Montreal PQ Canada Natl Taiwan Univ Dept Comp Sci & Informat Engn Taipei Taiwan

ISBN: (纸本)9781450392037

The storage of multi-media data can benefit from the advancements in general-purpose lossless compression. The explosive growth of multi-media data volume in data centers demands a higher compression ratio and better compressors' run-time speed. However, recent deep-learning-based compressors with a high compression ratio usually build complicated dependencies on history symbols, leading to a long compression time. This paper investigates the behavior of historical symbols and finds an approximate order of importance. Namely, recent symbols have a substantially larger influence on the probability estimation of the next unknown symbol. This observation guides the designing of an interpretable structure for data compression, rather than learning implicitly from data like Recurrent Neural Network (RNN) and attention. Based on this observation, we disentangle the compression model into order learning and feature learning, which were fused in a large module in previous works. A parameterized ordered mask unit is established to learn the ordered importance of history symbols. A fast Multi-Layer Perceptron (MLP) network is designed for efficient feature learning. The proposed compressor can improve both compression performance and computational efficiency compared with transformer-based or RNN-based compressors. To further enhance computational efficiency, we propose a branch-MLP block to replace the original MLP layer. This block reduces the parameters and the FLOPs of the original MLP to a half, without sacrificing compression performance. Experiments on multi-media data demonstrate that our model improves the compression ratio by 10% on average across data domains while accelerating compression speed by 100% compared with the state-of-the-art. The source code and appendix are released at https://***/mynotwo/compressor_via_ simple_and_scalable_***.

关键词： multi-layer perceptron general-purpose compressor lossless data compression neural networks ordered importance computational efficient

来源：评论

学校读者我要写书评

暂无评论

Binary Forward-Only Algorithms

IEEE Design and Test

引用

IEEE Design and Test 2025年

作者： Huang, Baichuan Aminifar, Amir

Today, the overwhelming majority of Internet of Things (IoT) and mobile edge devices have extreme resource limitations, e.g., in terms of computing, memory, and energy. As a result, training Deep Neural Networks (DNNs) using the complex Backpropagation (BP) algorithm on such edge devices presents a major challenge. Forward-only algorithms have emerged as more computation- and memory-efficient alternatives without the requirement for backward passes. In this paper, we investigate binarizing state-of-the-art forward-only algorithms, which are applied to the forward passes of PEPITA, FF, and CwComp. We evaluate these forward-only algorithms with binarization and demonstrate that weight-only binarization may be up to 31× more efficient in terms of memory, with minor degradation in classification performance. Furthermore, we investigate and compare BP and forward-only algorithms in terms of binarization, finding that PEPITA and FF are more vulnerable to binary activations. The code is available at https://***/whubaichuan/BinaryFO. © 2013 IEEE.

关键词： binary neural network computational efficient forward-only IoT memory efficient TinyML

来源：评论

学校读者我要写书评

暂无评论

Optimizing Electrode Configurations for Wearable EEG Seizure Detection Using Machine Learning

引用

SENSORS 2023年第13期23卷 5805页

作者： Gelbard-Sagiv, Hagar Pardo, Snir Getter, Nir Guendelman, Miriam Benninger, Felix Kraus, Dror Shriki, Oren Ben-Sasson, Shay NeuroHelp Ltd IL-5252181 Ramat Gan Israel Ben Gurion Univ Negev Dept Cognit & Brain Sci IL-8410501 Beer Sheva Israel Beilinson Med Ctr Rabin Med Ctr Dept Neurol IL-4941492 Petah Tiqwa Israel Tel Aviv Univ Sackler Fac Med IL-6997801 Tel Aviv Israel Schneider Childrens Med Ctr Israel Dept Pediat Neurol IL-4920235 Petah Tiqwa Israel

Epilepsy, a prevalent neurological disorder, profoundly affects patients' quality of life due to the unpredictable nature of seizures. The development of a reliable and user-friendly wearable EEG system capable of detecting and predicting seizures has the potential to revolutionize epilepsy care. However, optimizing electrode configurations for such systems, which is crucial for balancing accuracy and practicality, remains to be explored. This study addresses this gap by developing a systematic approach to optimize electrode configurations for a seizure detection machine-learning algorithm. Our approach was applied to an extensive database of prolonged annotated EEG recordings from 158 epilepsy patients. Multiple electrode configurations ranging from one to eighteen were assessed to determine the optimal number of electrodes. Results indicated that the performance was initially maintained as the number of electrodes decreased, but a drop in performance was found to have occurred at around eight electrodes. Subsequently, a comprehensive analysis of all eight-electrode configurations was conducted using a computationally intensive workflow to identify the optimal configurations. This approach can inform the mechanical design process of an EEG system that balances seizure detection accuracy with the ease of use and portability. Additionally, this framework holds potential for optimizing hardware in other machine learning applications. The study presents a significant step towards the development of an efficient wearable EEG system for seizure detection.

关键词： seizure detection wearable EEG machine learning continuous EEG monitoring electrode configuration optimization computational efficient metric adjustment

来源：评论

学校读者我要写书评

暂无评论

High resolution for software-defined GPS-based SAR imaging using waveform-modulated range-compressed pulse: field experimental demonstration

引用

SIGNAL IMAGE AND VIDEO PROCESSING 2020年第4期14卷 655-663页

作者： Zheng, Yu Zhang, Zhuxian Feng, Lu Huang, Caixia Zhu, Peidong Wu, Peng Changsha Univ Coll Elect Commun & Elect Engn Hongshan Rd 98 Changsha Peoples R China Hong Kong Polytech Univ Dept Land Surveying & Geoinformat Hong Kong Peoples R China Natl Univ Def Technol Coll Elect Sci Changsha 410073 Peoples R China Changsha Univ Coll Comp Engn & Appl Math Hongshan Rd 98 Changsha Peoples R China

This paper presents an experimental demonstration with respect to a high-range-resolution imaging scheme for software-defined receiver-based passive GPS-based SAR, and a signal processing scheme for imaging is proposed correspondingly. In the proposed scheme, to reduce the computational complexity and high-memory requirement for receiver, GPS signals are down-converted to baseband for digitized collection at first. After performing range compression, the region of interest is selected based on the numbers of detected compressed pulses. Thereafter, the detected compressed pulses are up-sampled, modulated by a waveform with the frequency larger than bandwidth value and then spectrum-equalized for obtaining high-resolution range-compressed signals. The field results obtained from both land and ocean scenarios indicate that compared to the conventional signal processing scheme for imaging, indeed the proposed scheme can provide a significantly lower range ambiguity around the scene center;compared with the authors' previous related work (Zheng et al. in Sensors, 2017. 10.3390/s17071496), the proposed scheme is obviously more computationally efficient.

关键词： GPS-SAR Range resolution Field experimental demonstration computational efficient

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：