Search

In the sintering process, it is difficult to obtain the key quality variables in real time, so there is lack of real-time information to guide the production process. Furthermore, these labeled data are too few, resulting in poor performance of conventional soft sensor models. Therefore, a novel semi-supervised dynamic feature extraction framework (SS-DTFEE) based on sequence pre-training and fine-tuning is proposed in this paper. Firstly, based on the DTFEE model, the time features of the sequences are extended and extracted. Secondly, a novel weighted bidirectional LSTM unit (BiLSTM) is designed to extract the latent variables of original sequence data. Based on improved BiLSTM, an encoder-decoder model is designed as a pre-training model with unsupervised learning to obtain the hidden information in the process. Next, through model migration and fine-tuning strategy, the prediction performance of labeled datasets is improved. The proposed method is applied in the actual sintering process to estimate the FeO content, which shows a significant improvement of the prediction accuracy, compared to traditional methods.

This paper presents a novel state-feedback control scheme for the tracking control of a class of multi-input multioutput continuous-time nonlinear systems with unknown dynamics and bounded disturbances. First, the control law consisting of the robust integral of a neural network (NN) output plus sign of the tracking error feedback multiplied with an adaptive gain is introduced. The NN in the control law learns the system dynamics in an online manner, while the NN residual reconstruction errors and the bounded disturbances are overcome by the error sign signal. Since both of the NN output and the error sign signal are included in the integral, the continuity of the control input is ensured. The controller structure and the NN weight update law are novel in contrast with the previous effort, and the semiglobal asymptotic tracking performance is still guaranteed by using the Lyapunov analysis. In addition, the NN weights and all other signals are proved to be bounded simultaneously. The proposed approach also relaxes the need for the upper bounds of certain terms, which are usually required in the previous designs. Finally, the theoretical results are substantiated with simulations.

It is difficult to directly coregister the 3-D fluorescence molecular tomography (FMT) image of a small tumor in a mouse whose maximal diameter is only a few millimeters with a larger CT image of the entire animal that spans about 10 cm. This paper proposes a new method to register 2-D flat and 3-D CT image first to facilitate the registration between small 3-D FMT images and large 3-D CT images. A novel algorithm combining differential evolution and improved simplex method for the registration between the 2-D flat and 3-D CT images is introduced and validated with simulated images and real images of mice. The visualization of the alignment of the 3-D FMT and CT image through 2-D registration shows promising results.

Wireless Sensor Network (WSN) is increasingly popular in the field of micro-environmental monitoring due to its promising capability. However, most systems using WSN for environmental monitoring reported in the literature are developed for specific applications without functions for exploiting user's data processing methods. In this paper, a new system is designed in detail to perform micro-environmental monitoring taking the advantages of the WSN. The application-oriented hardware working style is designed, and the system platform for data acquisition, validation, processing and visualization is systematically presented. Several strategies are proposed to guarantee the system capability in terms of extracting useful information, visualizing events to their authentic time are also described. Moreover, a web-based surveillance subsystem is presented for remote control and monitoring. In addition, the system is extensible for engineers to carry their own data analysis algorithms. Experimental results are to show the path reliability and real-time characteristics, and to display the feasibility and applicability of the developed system into practical deployment.

How to support real-time traffic of aperiodic messages using WorldFIP is investigated in this paper. Transmission of aperiodic message within WorldFIP is a very complicated procedure because it is concerned with left time in a microcycle and must successively experience three different stages, indicating, identifying, and transmitting aperiodic message. This paper presents an integrated approach to analyze the worst-case responding time (WCRT) of aperiodic message, through integrating transmission of periodic message and these three stages of aperiodic message together. To decrease pessimistic factors in WCRT of aperiodic message, priority of aperiodic message is considered. Detail analysis of the WCRT of aperiodic message indicates that the current scheduling method cannot effectively utilize priority to guarantee important aperiodic messages. Then improved scheduling methods, one adding priority and another adjusting the schedule table, are proposed. The improved methods are finally validated by case studies.

We study the problem of minimum-number full-view area coverage in camera sensor networks, i.e., how to select the minimum number of camera sensors to guarantee the full-view coverage of a given region. Full-view area coverage is challenging because the full-view coverage of a 2-D continuous domain has to be considered. To tackle this challenge, we first study the intrinsic geometric relationship between the full-view area coverage and the full-view point coverage and prove that the full-view area coverage can be guaranteed, as long as a selected full-view ensuring set of points is full-view covered. This leads to a significant dimension reduction for the full-view area coverage problem. Next, we prove that the minimum-number full-view point coverage is NP-hard and propose two approximation algorithms to solve it from two different perspectives, respectively: 1) By introducing a full-view coverage ratio function, we quantify the “contribution” of each camera sensor to the full-view coverage through which we transform the full-view point coverage into a submodular set cover problem and propose a greedy algorithm (GA); and 2) by studying the geometric relationship between the full-view coverage and the traditional coverage, we propose a set-cover-based algorithm (SCA). We analyze the performance of these two approximation algorithms and characterize their approximation ratios. Furthermore, we devise two distributed algorithms that obtain the same approximation ratios as GA and SCA, respectively. Finally, we provide extensive simulation results to validate our analysis. [ABSTRACT FROM PUBLISHER]

This paper proposes a novel cooperative charging strategy for a smart charging station in the dynamic electricity pricing environment, which helps electric vehicles (EVs) to economically accomplish the charging task by the given deadlines. This strategy allows EVs to share their battery-stored energy with each other under the coordination of an aggregator, so that more flexibility is given to the aggregator for better scheduling. Mathematically, the scheduling problem is formulated as a constrained mixed-integer linear program (MILP) to capture the discrete nature of the battery states, i.e., charging, idle and discharging. Then, an efficient algorithm is proposed to solve the MILP by means of dual decomposition and Benders decomposition. At last, the algorithm can be implemented in a distributed fashion, which makes it scalable and thus suitable for large-scale scheduling problems. Numerical results validate our theoretical analysis. [ABSTRACT FROM PUBLISHER]

Multiple channels in Wireless Sensor Networks (WSNs) are often exploited to support parallel transmission and to reduce interference. However, the extra overhead posed by the multi-channel usage coordination dramatically challenges the energy-constrained WSNs. In this paper, we propose a Regret Matching based Channel Assignment algorithm (RMCA) to address this challenge, in which each sensor node updates its choice of channels according to the historical record of these channels’ performance to reduce interference. The advantage of RMCA is that it is highly distributed and requires very limited information exchange among sensor nodes. It is proved that RMCA converges almost surely to the set of correlated equilibrium. Moreover, RMCA can adapt the channel assignment among sensor nodes to the time-variant flows and network topology. Simulations show that RMCA achieves better network performance in terms of both delivery ratio and packet latency than CONTROL refid="ref1" , MMSN refid="ref2"/ and randomized CSMA. In addition, real hardware experiments are conducted to demonstrate that RMCA is easy to be implemented and performs better. [ABSTRACT FROM PUBLISHER]

In this paper, the use of multiple variable spaces is proposed for monitoring modern industrial processes where data for a large number of process variables may be collected from different sources to reveal different characteristics. The easiest method of modeling a process is to treat all variables in a single data space, but then the information inherent in different types of variables would be mixed together and there would be no local view of each variable space. An extended algorithm based on the concept of total projection to latent structures, which we call multispace T-PLS (MsT-PLS), is thus developed to treat variables in multiple data spaces. Multiple variable spaces that are separated from the measurement space are composed of different sets of process variables measured at the same time and responsible for the same response data. Using the proposed algorithm, the relationships among multiple variable spaces are studied under the supervision of quality characteristics. Thus, comprehensive information decomposition is obtained in each variable space, which can be separated into four systematic subspaces in response to the cross-space common and specific process variability and one final residual subspace. The theoretical support for MsT-PLS is analyzed in detail and its statistical characteristics are compared with those of single-space T-PLS (SsT-PLS) algorithm. A process monitoring strategy is developed based on the MsT-PLS subspace decomposition result and applied to the Tennessee Eastman process for illustration purposes. [ABSTRACT FROM AUTHOR]

Time-varying fault characteristics have not yet beenaddressedby conventional fault-reconstruction-based modeling methods, whichcould affect fault diagnosis performance. In the present work, themultiple-time-region (MTR) nature, that is, the multiplicity of faultcharacteristics along with the process evolution, is proposed andefficiently analyzed for fault diagnosis. First, an automatic time-region-divisionalgorithm is developed that can partition the whole fault processinto different local regions according to the changes in fault characteristics.Different local fault characteristics are thus analyzed by buildingdifferent representative fault feature models in multiple time regions.Following the changing relationships between the fault and normaloperation statuses, different fault reconstruction actions are finallytaken in different time regions. By a proper time-region division,the proposed method can better model the time-varying fault behaviorsand capture the different fault-to-normal reconstruction relationshipsfor fault diagnosis. The feasibility and performance of the proposedmethod are illustrated with the Tennessee Eastman process, revealingenhanced fault understanding and improved fault diagnosis performance. [ABSTRACT FROM AUTHOR]

As sensors are energy constrained devices, one challenge in wireless sensor networks (WSNs) is to guarantee coverage and meanwhile maximize network lifetime. In this paper, we leverage prediction to solve this challenging problem, by exploiting temporal-spatial correlations among sensory data. The basic idea lies in that a sensor node can be turned off safely when its sensory information can be inferred through some prediction methods, like Bayesian inference. We adopt the concept of entropy in information theory to evaluate the information uncertainty about the region of interest (RoI). We formulate the problem as a minimum weight submodular set cover problem, which is known to be NP hard. To address this problem, an efficient centralized truncated greedy algorithm (TGA) is proposed. We prove the performance guarantee of TGA in terms of the ratio of aggregate weight obtained by TGA to that by the optimal algorithm. Considering the decentralization nature of WSNs, we further present a distributed version of TGA, denoted as DTGA, which can obtain the same solution as TGA. The implementation issues such as network connectivity and communication cost are extensively discussed. We perform real data experiments as well as simulations to demonstrate the advantage of DTGA over the only existing competing algorithm [1] and the impacts of different parameters associated with data correlations on the network lifetime. [ABSTRACT FROM AUTHOR]

Abstract: An adaptive inverse control algorithm is proposed by combining fast online support vector machine regression (SVR) algorithm with straight inverse control algorithm. Because training speed of standard online SVR algorithm is very slow, a kernel cache-based method is developed to accelerate the standard algorithm and a new fast online SVR algorithm is obtained. Then the new algorithm is applied in straight inverse control for constructing the inverse model of controlled system online, and output errors of the system are used to control online SVR algorithm, which made the whole control system a closed-loop one. Simulation results show that the new algorithm has good control performance. [Copyright &y& Elsevier]

Abstract: For the design of a more efficient congestion control algorithm to optimize the resource allocation in Ad Hoc networks, the notion of the link’s interference set is introduced to represent the contention relationship among flows in Ad Hoc networks, which is the fundamental difference from the contention relationship among flows in Internet. And the congestion control problem is formulated as a nonlinear optimization problem with the capacity constraint of the link’s interference set. Then, based on the dual decomposition method, this problem is decomposed into many subproblems, which can be solved independently in a parallel fashion at each source/link, and builds the Dual Decomposition based Optimal and Fair Congestion Control framework () to coordinate their source allocation among flows. Considering the synchronous implementation is difficult in realistic environment, the asynchronous implementation of is presented. To deploy the practical protocol of , three deployment techniques are proposed: Queue Size Estimation method (QSE), Neighbor Set Approximation method (NSA), and Extensive HELLO message-based Piggybacking (EHP).The MATLAB-based simulation results illustrate that can approach the globally optimal solution even in asynchronous environment. The NS2-based simulation results demonstrate that outperforms default TCP and ATCP to achieve an efficient and fair resource allocation in Ad Hoc networks under three scenarios, including chain topology, grid topology, and random topology. [Copyright &y& Elsevier]