Output-relevant Variational autoencoder for Just-in-time soft sensor modeling with missing data

Main challenges for developing data-based models lie in the existence of high-dimensional and possibly missing observations that exist in stored data from industry process. Variational autoencoder (VAE) as one of the deep learning methods has been applied for extracting useful information or features from high-dimensional dataset. Considering that existing VAE is unsupervised, an output-relevant VAE is proposed for extracting output-relevant features in this work. By using correlation between process variables, different weight is correspondingly assigned to each input variable. With symmetric Kullback–Leibler (SKL) divergence, the similarity is evaluated between the stored samples and a query sample. According to the values of the SKL divergence, data relevant for modeling are selected. Subsequently, Gaussian process regression (GPR) is utilized to establish a model between the input and the corresponding output at the query sample. In addition, owing to the common existence of missing data in output data set, the parameters and missing data in the GPR are estimated simultaneously. A practical debutanizer industrial process is utilized to illustrate the effectiveness of the proposed method.     

A bi-layer optimization approach for a hybrid flow shop scheduling problem involving controllable processing times in the steelmaking industry

A steelmaking-continuous casting (SCC) scheduling problem is an example of complex hybrid flow shop scheduling problem (HFSSP) with a strong industrial background. This paper investigates the SCC scheduling problem that involves controllable processing times (CPT) with multiple objectives concerning the total waiting time, earliness/tardiness and adjusting cost. The SCC scheduling problem with CPT is seldom discussed in the existing literature. This study is motivated by the practical situation of a large integrated steel company in which the just-in-time (JIT) and cost-cutting production strategy have become a significant concern. To address this complex HFSSP, the scheduling problem is decomposed into two subproblems: a parallel machine scheduling problem (PMSP) in the last stage and an HFSSP in the upstream stages. First, a hybrid differential evolution (HDE) algorithm combined with a variable neighborhood decomposition search (VNDS) is proposed for the former subproblem. Second, an iterative backward list scheduling (IBLS) algorithm is presented to solve the latter subproblem. The effectiveness of this bi-layer optimization approach is verified by computational experiments on well-designed and real-world scheduling instances. This study provides a new perspective on modeling and solving practical SCC scheduling problems.     

A design of a robust discrete-time controller

In this paper, we proposed a robust discrete-time controller. This control system, which is derived from the idea of the normalized plant, does not include plant parameters. Thus, we obtain a control system independent of plant parameters and that has the same structure as a conventional optimal servo control system. Simulation and experimental results show that the proposed method is fairly robust to plant parameter variations and external disturbances.     

New Paradigm of Data-Driven Smart Customisation through Digital Twin

Big data is one of the most important resources for the promotion of smart customisation. With access to data from multiple sources, manufacturers can provide on-demand and customised products. However, existing research of smart customisation has focused on data generated from the physical world, not virtual models. As physical data is constrained by what has already occurred, it is limited in the identification of new areas to improve customer satisfaction. A new technology called digital twin aims to achieve this integration of physical and virtual entities. Incorporation of digital twin into the paradigm of existing data-driven smart customisation will make the process more responsive, adaptable and predictive. This paper presents a new framework of data-driven smart customisation augmented by digital twin. The new framework aims to facilitate improved collaboration of all stakeholders in the customisation process. A case study of the elevator industry illustrates the efficacy of the proposed framework.     

An online multiple object tracker based on structure keeper net

We propose a novel online multiple object tracker taking structure information into account. State-of-the-art multi-object tracking (MOT) approaches commonly focus on discriminative appearance features, while neglect in different levels structure information and the core of data association. Addressing this, we design a new tracker fully exploiting structure information and encoding such information into the cost function of the graph matching model. Firstly, a new measurement is proposed to compare the structure similarity of two graphs whose nodes are equal. With this measurement, we define a complete matching which performs association in high efficiency. Secondly, for incomplete matching scenarios, a structure keeper net (SKnet) is designed to adaptively establish the graph for matching. Finally, we conduct extensive experiments on benchmarks including MOT2015 and MOT17. The results demonstrate the competitiveness and practicability of our tracker.

     

基于强跟踪滤波器的开关变换器故障诊断

Buck变换器又称降压变换器、串联开关稳压电源、三端开关型降压稳压器。广泛用于计算机、家用电器、控制系统和通信系统中的电源几乎都是开关电源,具有超快速负载动态响应、高功率变换效率、高功率密度、低电压、大电流的特性.开关变换器作为它的主要电路,故障率较高,对它故障诊断研究是有意义的。基于强跟踪滤波理论,给出了一种开关变换器电路故障实时诊断的一种方法。该方法以建立一种开关变换器电路状态空间模型为基础,利用强跟踪滤波器对电路状态及元件参数进行估计,当元件参数发生软、硬型故障时,根据强跟踪滤波器对元件参数的跟踪结果及修正的Bayes分类算法,可实时诊断开关变换器电路中的元件故障。仿真结果证明了该方法的有效性。     

PUF-enhanced offline RFID security and privacy

RFID (Radio Frequency IDentification) based communication solutions have been widely used nowadays for mobile environments such as access control for secure system, ticketing systems for transportation, and sport events. These systems usually depend on readers that are not continuously connected to a secure backend system. Thus, the readers should be able to perform their duties even in offline mode, which generally requires the management by the readers of the susceptible data. The use of RFID may cause several security and privacy issues such as traceability of tag owner, malicious eavesdropping and cloning of tags. Besides, when a reader is compromised by an adversary, the solution to resolve these issues getting worse. In order to handle these issues, several RFID authentication protocols have been recently proposed; but almost none of them provide strong privacy for the tag owner. On the other hand, several frameworks have been proposed to analyze the security and privacy but none of them consider offline RFID system.Motivated by this need, in this paper, we first revisit Vaudenay's model, extend it by considering offline RFID system and introduce the notion of compromise reader attacks. Then, we propose an efficient RFID mutual authentication protocol. Our protocol is based on the use of physically unclonable functions (PUFs) which provide cost-efficient means to the fingerprint chips based on their physical properties. We prove that our protocol provides destructive privacy for tag owner even against reader attacks.     

A multi-document summarization system based on statistics and linguistic treatment

The massive quantity of data available today in the Internet has reached such a huge volume that it has become humanly unfeasible to efficiently sieve useful information from it. One solution to this problem is offered by using text summarization techniques. Text summarization, the process of automatically creating a shorter version of one or more text documents, is an important way of finding relevant information in large text libraries or in the Internet. This paper presents a multi-document summarization system that concisely extracts the main aspects of a set of documents, trying to avoid the typical problems of this type of summarization: information redundancy and diversity. Such a purpose is achieved through a new sentence clustering algorithm based on a graph model that makes use of statistic similarities and linguistic treatment. The DUC 2002 dataset was used to assess the performance of the proposed system, surpassing DUC competitors by a 50% margin of f-measure, in the best case.