Future embedded systems demand increasingly more computation performance, which can only be provided by exploiting parallelism in real-time applications. Due to scheduling and scalability issues, parallelism still is ...
详细信息
distributed estimation over wireless sensor networks (WSNs) has been used to obtain the parameters of interest with reduced resource consumption, hence gained importance in system modeling and control applications. Un...
详细信息
ISBN:
(纸本)9781538647271
distributed estimation over wireless sensor networks (WSNs) has been used to obtain the parameters of interest with reduced resource consumption, hence gained importance in system modeling and control applications. Unlike least-squares and fusion-center based approaches, distributed signal processing is competent in real-time applications. In this article, Volterra-Laguerre model and Wiener model are identified in a distributed manner through WSNs for modeling of nonlinear systems. A block-structured Wiener model has been widely used as it is characterized by a small number of parameters, but can only model specific nonlinearities. A generalized Volterra model over Wiener model can approximate any nonlinear system to a desired precision but has increased parameter complexity. By expanding nonlinear Volterra kernels with orthogonal Laguerre functions, the parameter complexity is reduced significantly. A distributed recursive algorithm for the identification of abovementioned nonlinear models is designed by minimizing the quadratic prediction error. The algorithm reformulates model identification framework into multiple constrained separable subtasks. These subtasks are optimized using a powerful method called alternating direction method of multipliers. Simulation results for an infinite-order and a 2nd-order nonlinear systems are obtained under the influence of process noise and are compared with the results of non-cooperative estimation showing the superiority of the proposed algorithm.
We illustrate a devised and implemented data acquisition system (DAQ) for a magnetic positioning system (MPS) that is currently under development. This system aims to track position and attitude of an active transmitt...
详细信息
ISBN:
(纸本)9781538644461
We illustrate a devised and implemented data acquisition system (DAQ) for a magnetic positioning system (MPS) that is currently under development. This system aims to track position and attitude of an active transmitting coil (TX) supplied with a sinusoidal current, whose generated magnetic field induces tensions on an array of passive receiving coils (RX). The DAQ system has to acquire voltages at all RX coils. These signals are then processed to estimate the TX coordinates, according to a mathematical model in order. In order to track the TX in real-time with a good resolution, voltages have to be measured simultaneously on all RXs. To this aim, we opted for a distributed architecture of microcontroller units (MCU). Each selected MCU has four analog-to-digital converters (ADC) that can work in parallel. Moreover multiple MCUs can be triggered simultaneously by a single MCU in a master-slave configuration. We used MCUs with a fast dual-core CPU. Each unit can directly process its own acquired signals. Then all data are sent to the master MCU, which estimates the coordinates of the TX. According to a preliminary analisys, this tracking system should achieve more than fifty coordinates measurements per second.
Some problems of a real-timedistributed computer system testing are considered. The authors approach suggests the introduction of redundancy into the system being diagnosed, which in essence is an event model of the ...
详细信息
Some problems of a real-timedistributed computer system testing are considered. The authors approach suggests the introduction of redundancy into the system being diagnosed, which in essence is an event model of the DS. The model is a dynamical periodically time-varying system, linear in a binary field. Based on this model rational procedure for designing the testing of a distributed system is proposed. (C) 2018, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Coding for distributed computing supports low-latency computation by relieving the burden of straggling workers. While most existing works assume a simple master-worker model, we consider a hierarchical computational ...
详细信息
ISBN:
(纸本)9781538647813
Coding for distributed computing supports low-latency computation by relieving the burden of straggling workers. While most existing works assume a simple master-worker model, we consider a hierarchical computational structure consisting of groups of workers, motivated by the need to reflect the architectures of real-world distributed computing systems. In this work, we propose a hierarchical coding scheme for this model, as well as analyze its decoding cost and expected computation time. Specifically, we first provide upper and lower bounds on the expected computing time of the proposed scheme. We also show that our scheme enables efficient parallel decoding, thus reducing decoding costs by orders of magnitude over non-hierarchical schemes. When considering both decoding cost and computing time, the proposed hierarchical coding is shown to outperform existing schemes in many practical scenarios.
Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex realsystems in different fields. Finding an optimal system design requires performing a large sweep over the pa...
详细信息
Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex realsystems in different fields. Finding an optimal system design requires performing a large sweep over the parameter space in an organized way. Hence, the model optimization process is extremely demanding from a computational point of view, as it requires careful, time-consuming, complex orchestration of coordinated executions. In this paper, we present the design of SOF (Simulation Optimization and exploration Framework in the cloud), a framework which exploits the computing power of a cloud computational environment in order to carry out effective and efficient simulation optimization strategies. SOF offers several attractive features. Firstly, SOF requires "zero configuration", as it does not require any additional software installed on the remote node;only standard Apache Hadoop and SSH access are sufficient. Secondly, SOF is transparent to the user, since the user is totally unaware that the system operates on a distributed environment. Finally, SOF is highly customizable and programmable, since it enables the running of different simulation optimization scenarios using diverse programming languages - provided that the hosting platform supports them - and different simulation toolkits, as developed by the modeler. The tool has been fully developed and is available on a public repository 1 under the terms of the open source Apache License. It has been tested and validated on several private platforms, such as a dedicated cluster of workstations, as well as on public platforms, including the Hortonworks Data Platform and Amazon Web Services Elastic MapReduce solution. (C) 2017 Elsevier B.V. All rights reserved.
In the paper the method of fuzzy clustering task for multi-variate short time series with unevenly distributed observations is proposed. Proposed method allows to process the time series both in batch mode and sequent...
详细信息
ISBN:
(纸本)9783319668277;9783319668260
In the paper the method of fuzzy clustering task for multi-variate short time series with unevenly distributed observations is proposed. Proposed method allows to process the time series both in batch mode and sequential on-line mode. In the first case we can use the matrix modification of fuzzy C-means method, and in second case we can use the matrix modification of neuro-fuzzy network by T. Kohonen, which is learned using the rule "Winner takes more". Proposed fuzzy clustering algorithms are enough simple in computational implementation and can be used for solving of wide class of Big Data and Data Stream Mining problems. The effectiveness of proposed approach is confirmed by many experiments based on real data sets.
We have developed a web-based feedback system that students can use to report difficult, easy, engaging, and boring sections of a lecture in real-time. Such feedback can identify potentially problematic lecture conten...
详细信息
We have developed a web-based feedback system that students can use to report difficult, easy, engaging, and boring sections of a lecture in real-time. Such feedback can identify potentially problematic lecture content, track cognitive-affective dynamics in a classroom, and assist instructors in retrospective self-evaluation. We use a mixed-method approach within a design-based research (DBR) framework. In this paper, we discuss initial development and implementation of the feedback system, which constitutes the first cycle of DBR in the project. We discuss potential use-cases of such data and follow the DBR framework to identify strengths and weaknesses in the current prototype and its implementation. Based on such an analysis we come up with a list of modifications for the next cycle of DBR.
This book presents an integrated collection of representative approaches for scaling up machine learning and data mining methods on parallel and distributed computing platforms. Demand for parallelizing learning algor...
ISBN:
(纸本)1108461743
This book presents an integrated collection of representative approaches for scaling up machine learning and data mining methods on parallel and distributed computing platforms. Demand for parallelizing learning algorithms is highly task-specific: in some settings it is driven by the enormous dataset sizes, in others by model complexity or by real-time performance requirements. Making task-appropriate algorithm and platform choices for large-scale machine learning requires understanding the benefits, trade-offs, and constraints of the available options. Solutions presented in the book cover a range of parallelization platforms from FPGAs and GPUs to multi-core systems and commodity clusters, concurrent programming frameworks including CUDA, MPI, MapReduce, and DryadLINQ, and learning settings (supervised, unsupervised, semi-supervised, and online learning). Extensive coverage of parallelization of boosted trees, SVMs, spectral clustering, belief propagation and other popular learning algorithms and deep dives into several applications make the book equally useful for researchers, students, and practitioners.
暂无评论