检索结果-内蒙古大学图书馆

IEEE International conference on Solid-State Circuits (ISSCC)

作者： Naveen Verma Princeton University Princeton NJ

ISBN: (数字)9781509049400

ISBN: (纸本)9781538622278

Deep-learning systems have had profound impacts in a broad range of applications. However, it is important to remember that these represent only one class ofmachine learning. In this segment of the short course, we start by probing what the critical attributes are of deep learning, and what challenges in modeling andinference they solve. We then go on to consider the limitations of deep learning in emerging applications involving on-line learning (e.g. reinforcement learningwith embedded sensors) - namely the need for a large number of training instances and the need for very low energy. This motivates alternatives or extensions todeep learning, which make use of other forms of learning to enhance training and energy efficiency. Given the need for very low energy in many applications, weexplore how the statistical-learning can enable new hardware architectures, substantially overcoming the tradeoffs limiting conventional architectures for sensingand computation. Finally, having examined how algorithmic techniques can enhance systems, we look at how systems, and emerging technologies for sensing,can enhance algorithms. As an illustration, we consider how object-associated sensing, as enabled by IoT devices, has the potential to provide semantic structure,leading to features that can enhance the generalization of learning with simpler and easier-to-train models.

关键词： Tutorials Radio frequency Wireless communication Transceivers Solid state circuits Quantum computing Optical transmitters Computer architecture Wireless sensor networks Training

来源：评论

学校读者我要写书评

暂无评论

Set-Membership Information fusion for Multisensor Nonlinear Dynamic Systems 20

Set-Membership Information Fusion for Multisensor Nonlinear ...

引用

20th International conference on Information fusion (fusion)

作者： Wang, Zhiguo Shen, Xiaojing Zhu, Yunmin Sichuan Univ Sch Math Chengdu 610064 Sichuan Peoples R China

ISBN: (纸本)9780996452700

The set-membership information fusion problem is investigated for general multisensor nonlinear dynamic systems. Compared with linear dynamic systems and point estimation fusion in mean squared error sense, it is a more challenging nonconvex optimization problem. Usually, to solve this problem, people try to find an efficient or heuristic fusion algorithm. It is no doubt that an analytical fusion formula should be much significant for raising accuracy and reducing computational burden. However, since it is a more complicated than the convex quadratic optimization problem for linear point estimation fusion, it is not easy to get the analytical fusion formula. In order to overcome the difficulty of this problem, two popular fusion architectures are considered: centralized and distributed set-membership information fusion. Firstly, both of them can be converted into a semidefinite programming problem which can be efficiently computed, respectively. Secondly, their analytical solutions can be derived surprisingly by using decoupling technique. It is very interesting that they are quite similar in form to the classic information filter. In the two analytical fusion formulae, the information of each sensor can be clearly characterized, and the knowledge of the correlation among measurement noises across sensors are not required. Finally, multi-algorithm fusion is used to minimize the size of the state bounding ellipsoid by complementary advantages of multiple parallel algorithms. A typical numerical example in target tracking demonstrates the effectiveness of the centralized, distributed, and multi-algorithm set-membership fusion algorithms. In particular, it shows that multi-algorithm fusion performs better than the centralized and distributed fusion.

关键词： Nonlinear dynamic systems multisensor fusion target tracking unknown but bounded noise set-membership filter

来源：评论

学校读者我要写书评

暂无评论

量测不一致性条件下多传感器融合算法研究

引用

现代制造工程 2018年第8期 123-127,47页

作者：徐建亮顾怡红魏小华衢州职业技术学院机电工程学院衢州324000 浙江大学机械工程学院杭州310027

由于传感器提供的抽象数据总是受到测量过程中不确定性因素的影响,多传感器利用多源数据融合算法,消除传感器间可能存在的虚假或错误数据,以降低其不确定性。提出一种基于粒子滤波算法(PFA)的改进贝叶斯方法(IBA),其中基于粒子滤波有3... 详细信息

由于传感器提供的抽象数据总是受到测量过程中不确定性因素的影响,多传感器利用多源数据融合算法,消除传感器间可能存在的虚假或错误数据,以降低其不确定性。提出一种基于粒子滤波算法(PFA)的改进贝叶斯方法(IBA),其中基于粒子滤波有3种算法结构,即:前置粒子滤波算法(P-IBA)、后置粒子滤波算法(IBA-P)和前后置粒子滤波算法(P-IBA-P)。通过三维点云数据分析实验表明,P-IBA具有最小的CPU运行时间t,P-IBA-P比IBA-P拥有最小的评价函数值EF,可见,在P-IBA-P和IBA-P两种算法中,由于粒子滤波器的存在且起到决定性的作用,使得基于粒子滤波的改进贝叶斯算法(P-IBA-P)误差最小,精度最高,融合效果最好。改进的贝叶斯方法与粒子滤波相结合的技术,有助于处理集中式和分散式融合架构的数据不一致问题,在对测量时间要求不高的情况下,使用P-IBA-P算法的测量效果最好。

关键词：多传感器融合改进贝叶斯方法粒子滤波算法

来源：评论

学校读者我要写书评

暂无评论

Multisensor Data fusion for Human Activities Classification and Fall Detection 16

Multisensor Data Fusion for Human Activities Classification ...

引用

16th IEEE sensorS conference

作者： Li, Haobo Shrestha, Aman Fioranelli, Francesco Le Kernec, Julien Heidari, Hadi Pepa, Matteo Cippitelli, Enea Gambi, Ennio Spinsante, Susanna Univ Glasgow Sch Engn Glasgow Lanark Scotland Univ Politecn Marche Dept Informat Engn Ancona Italy

ISBN: (纸本)9781509010127

Significant research exists on the use of wearable sensors in the context of assisted living for activities recognition and fall detection, whereas radar sensors have been studied only recently in this domain. This paper approaches the performance limitation of using individual sensors, especially for classification of similar activities, by implementing information fusion of features extracted from experimental data collected by different sensors, namely a tri-axial accelerometer, a micro-Doppler radar, and a depth camera. Preliminary results confirm that combining information from heterogeneous sensors improves the overall performance of the system. The classification accuracy attained by means of this fusion approach improves by 11.2% compared to radar-only use, and by 16.9% compared to the accelerometer. Furthermore, adding features extracted from a RGB-D Kinect sensor, the overall classification accuracy increases up to 91.3%.

关键词： Accelerometer Radar sensor Depth camera human activities classification fall detection machine learning data fusion

来源：评论

学校读者我要写书评

暂无评论

Data fusion Utilization for Distributed Target Detection with Tree Topology 20

Data Fusion Utilization for Distributed Target Detection wit...

引用

20th International conference on Information fusion (fusion)

作者： Luo, Junhai Ni, Jing Wu, Qi Fan, Liying Univ Elect Sci & Technol China Sch Elect Engn Chengdu Sichuan Peoples R China

ISBN: (纸本)9780996452700

Multi-sensor fusion has been extensively studied in information fusion field, and the distributed target detection is one of the most important applications in the multiple sensor detection theories. In this paper, a data fusion algorithm for target detection is proposed based on tree topology combined with the orderly full binary tree and we discuss the optimal threshold fusion rule problem. Different from the conventional tree topology, the sensors in our topology are well ordered and the sensors with highest signal amplitude are selected as the fusion center. Moreover, we derive a fusion decision rule which takes the channel noise into consideration based on this topology. By introducing the probability of error, we prove that with equal prior probability there is a concave function of the likelihood ratio threshold used in the sensor decision rule. In the case of minimizing the probability of error, the optimal threshold of each level can be obtained. Finally, the distributed detection performance of tree topology is analyzed and comparisons with other topologies are drawn. These results show that the detection performance improves as the level decreases and our fusion rule can achieve a higher probability of detection.

关键词： Estimation Field programmable gate arrays OFDM Modulation Filtration Clocks Signal processing algorithms

来源：评论

学校读者我要写书评

暂无评论

Confidence Field-based Temporal Alignment and Positioning for Vehicles Using Multiple sensors 86

Confidence Field-based Temporal Alignment and Positioning fo...

引用

86th IEEE Vehicular Technology conference (VTC-Fall)

作者： Sun, Jinlong Wu, Zhilu Yin, Zhendong Harbin Inst Technol Sch Elect & Informat Engn Harbin 150001 Heilongjiang Peoples R China

ISBN: (纸本)9781509059355

Various vehicle applications in the future will require reliable and accurate vehicle positioning techniques. Nowadays, hybrid schemes combining multiple sensors have been promising solutions for high precision positioning. However, positioning error can be remarkably affected by the temporal alignment and fusion algorithms in practice. In this paper, we propose a decentralized fusion structure containing an inertial navigation system (INS), a GPS receiver, a RFID reader, and an odometer. The update rates of the sensors are different, and the INS/GPS integration presents severe performance degradation in urban area. To achieve an effective alignment and fusion of the sensors, we propose a concept of confidence field to indicate the confidence levels of subsystems for changing driving environments. A confidence field-based alignment and fusion algorithm and its simplification are proposed when we use the weighted least squares curve method. Time biases of the sensors are also considered in local adaptive filters. Simulation results demonstrates the performance of the proposed scheme with the proposed algorithms, especially in GPS-denied environments.

关键词： vehicle positioning temporal alignment sensor fusion confidence level adaptive filter GPS outages

来源：评论

学校读者我要写书评

暂无评论

Real-Time Wireless UWB sensor Network for Person Monitoring 14

Real-Time Wireless UWB Sensor Network for Person Monitoring

引用

14th International conference on Telecommunications (ConTEL)

作者： Drutarovsky, Milos Kocur, Dusan Svecova, Maria Garcia, Nuno M. Tech Univ Kosice Dept Elect & Multimedia Commun Kosice Slovakia Univ Beira Interior Dept Comp Sci Covilha Portugal

ISBN: (纸本)9789531842235

Ultra-wide band (UWB) radar is a very perspective technology for short-range localization and tracking of moving persons. UWB radar sensor network (UWB-SN) employing the centralized data fusion method can significantly improve tracking capabilities of more people in complex building environments. In this paper we present real-time performing wireless UWB-SN hardware demonstrator. Its sensor nodes use M-sequence UWB radars front-end and low-cost ARM based quad-core microcomputer (ARM-MC) as a main signal processing block. The ARM-MC based on Raspberry Pi provides processing power for the preprocessing of received raw radar signals, algorithms for detection and estimation of target's coordinates, and finally compression of data sent to the data fusion center. Low-rate data streams (3600-6000 bits/s/node) of compressed target coordinates are sent from each sensor node to the data fusion center in the central node by using RF transceivers with integrated ARM microcontroller. The proposed UWB-SN uses wireless RF communication by using FSK modulation in free short range devices (SRD) 868-870 MHz frequency band concurrently with operation of the UWB radar front-end. Experimental testing confirmed real-time performance of developed UWB-SN hardware and acceptable precision of software algorithms implemented in 32-bit arithmetic. The introduced modular architecture of UWB-SN can be used for fast development and testing of new real-time people tracking applications.

关键词： M-sequence person tracking sensor network UWB radar

来源：评论

学校读者我要写书评

暂无评论

sensor fusion for Robot Control through Deep Reinforcement Learning

Sensor Fusion for Robot Control through Deep Reinforcement L...

引用

IEEE/RSJ International conference on Intelligent Robots and Systems (IROS) / Workshop on Machine Learning Methods for High-Level Cognitive Capabilities in Robotics

作者： Bohez, Steven Verbelen, Tim De Coninck, Elias Vankeirsbilck, Bert Simoens, Pieter Dhoedt, Bart Univ Ghent Imec IDLab Dept Informat Technol Ghent Belgium

ISBN: (纸本)9781538626825

Deep reinforcement learning is becoming increasingly popular for robot control algorithms, with the aim for a robot to self-learn useful feature representations from unstructured sensory input leading to the optimal actuation policy. In addition to sensors mounted on the robot, sensors might also be deployed in the environment, although these might need to be accessed via an unreliable wireless connection. In this paper, we demonstrate deep neural network architectures that are able to fuse information generated by multiple sensors and are robust to sensor failures at runtime. We evaluate our method on a search and pick task for a robot both in simulation and the real world.

关键词： Robot sensing systems Neural networks Machine learning Convolution Fuses Training

来源：评论

学校读者我要写书评

暂无评论

Consensus-based distributed sensor fusion over a network 1

Consensus-based distributed sensor fusion over a network

引用

1st Annual IEEE conference on Control Technology and applications, CCTA 2017

作者： Zhang, Xueji Hengster-Movrić, Kristian Šebek, Michael Desmet, Wim Faria, Cassio Department of Control Engineering Czech Technical University Prague Czech Republic Department of Mechanical Engineering KU Leuven Belgium Flanders Make Leuven3001 Belgium Siemens Industry Software NV Leuven3001 Belgium

ISBN: (纸本)9781509021826

This paper studies the distributed filtering of noisy measurements of one scalar quantity. The considered network is composed of two sets of nodes: sensing nodes which perform the measuring task and non-sensing nodes which mediate between the sensing nodes. Inspired by Bayesian sensor fusion, three consensus-based algorithms are proposed. In the algorithms, graph edge weights in effect are determined based on variances of measurements of all the sensors. Evolution dynamics of the expected values and covariances of the state estimates throughout the network is analyzed. Numerical simulations are given to examine the fusion performance of the proposed scheme. The results are fairly consistent with the analytical solution regarding the statistical properties of the steady-state state estimates in the network. © 2017 IEEE.

关键词： State estimation

来源：评论

学校读者我要写书评

暂无评论

An Experimental Performance Evaluation of the Orientation Accuracy of Four Nine-Axis MEMS Motion sensors 5

An Experimental Performance Evaluation of the Orientation Ac...

引用

5th International conference on Enterprise Systems (ES) - Industrial Digitalization by Enterprise Systems

作者： Lin, Zhirong Xiong, Yongsheng Dai, Houde Xia, Xuke Chinese Acad Sci Haixi Inst Quanzhou Inst Equipment Mfg Bolan Av Jinjiang 362200 Fujian Peoples R China

ISBN: (纸本)9781538609361

MEMS (Micro-Electro-Mechanical Systems) inertial motion sensors have greatly improved the intelligence of wearable devices, sports fitness equipment, and consumer electronics. To avoid the disadvantages of each discrete inertial sensor, the tendency is to combine and fuse several different types of inertial motion sensors;thus, consumer-grade inertial motion sensors even have the same performance with the industrial-grade sensors. This study presents an overview of multi-axis sensor fusion algorithms and their applications. Furthermore, three typical consumer-grade AHRS (attitude and heading reference system) were verified compared with an industrial-grade sensor module (MTi-300), and three-axis orientation outputs of these sensors were compared. Static and dynamic experiment results show that MTi-300 has the best orientation accuracy (static error=0.05 degrees, dynamic error=0.5 degrees) compared to the other three sensor modules. Moreover, BNO055 has a better dynamic performance while compared with the other two consumer-grade motion sensors.

关键词： Motion sensor MEMS Attitude accuracy Multi-sensor fusion

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：