Kenneth R. Currie

An effective p-median model considering production factors in machine cell/part family formation

This paper is concerned with machine cell/part family (MC/PF) formation in cellular manufacturing. Among the solution methodologies for MC/PF formation, the mathematical programming approach seeks to find the optimal number of machine cells and associated part families. The p-median model has been proven to be a powerful mathematical programming model for solving the MC/PF problem. To cluster machines and parts, a new lrmedian formulation is proposed considering real-world production factors such as the operation sequences and production volumes for parts. Unlike existing p-median models, which rely on the classic binary part-machine incidence matrix (PMIM), this p-median model adopts a new similarity coefficient based on production factors that critically impact the MC/PF formation. A production data based PMIM is used, where each non-binary entry indicates actual intra-cell or inter-cell flows to or from machines by parts. In addition, a new efficiency measure for evaluating the goodness of the solution is proposed. A computational comparison shows the effectiveness of the proposed p-median model in terms of the computation time and solution quality as compared with previous mathematical programming approaches.

Cognitive radio network as wireless sensor network (II): Security consideration

This is the second paper in a series of using cognitive radio network as wireless sensor network. The motivation of the paper is to push the convergence of radar and communication systems into a unified cognitive network. This paper studies this vision from a secure point of view. We propose two methods for robust spectrum sensing in the same framework of cognitive radio network. The first method is based on robust principal component analysis (PCA), to separate spectrum sensing results into the low rank signal matrix and the sparse attack matrix. Using sparse attack cancellation in least squares, the second method iteratively estimates the relative transmitted power of primary user under the threats of attackers. Then the relative transmitted power of primary user can be calculated from the recovered signal matrix. Both two methods can detect the sparse compromised cognitive radio nodes and effectively obtain the relative transmitted power.

An intelligent grouping algorithm for cellular manufacturing

Abstract The methodology presented in this paper will provide a means of identifying part families/machine cells using design and manufacturing characteristics simultaneously. The technique used is a self-organizing neural network called interative-activation and competition (IAC) which acts as a content-addressable memory. This neural network is used to define a similarity index of the pairwise comparisons of parts based on a variety of design and manufacturing characteristics. A bond energy algorithm partitions the matrix of part similarity indices to create part families and inferred from the part families are machine cells. A brief example will be examined as well as discussion of the results.

Cognitive radio network as wireless sensor network (I): Architecture, testbed, and experiment

This paper explores the vision of a dual-use sensing/communication system based on Cognitive Radio Network (CRN). The motivation of the paper is to push the convergence of sensing and communication systems into a unified cognitive network. The concept design of this sensing/communication system is presented along with potential functions and challenges. A through tree target detection using real data collected by CRN testbed is demonstrated. The CRN testbed will be built based on Rice WARP nodes. To further exploit the advantages of dynamic spectrum access and frequency diversity, multi-frequency signal, which is similar to OFDM signal, are employed to for detection experiment in the harsh radio environment. The experiment results illustrate the vision of employing a CRN as wireless sensor network.

Efficient p-median mathematical programming approaches to machine-part grouping in group technology manufacturing

This article develops efficient p-median mathematical formulations for solving the machine-part grouping problem (MPGP) in group technology manufacturing and compares the performance of the formulations with that of existing p-median ones. In spite of the successful applications of MPGP that have been reported in the literature, existing p-median formulations have been restricted to small to medium-sized MPGP since they attempt to find the optimal solution over the entire feasible region of the constraint set without any prior knowledge about the median machines. Our formulations lead to rapid implementation of the model by introducing the idea of a candidate set of median machines, which consists of the machines that have a high possibility of serving as medians or seed machines for grouping. The candidate set of median machines plays the role of medians known in advance and enables the model to be implemented with the feasible region of a reduced constraint set. Furthermore, our alternative formulation can attack large-size MPGPs efficiently. Computational results show the comparative advantage of the formulations in terms of computation time and solution quality over existing p-median ones.

Development of an expert system for scheduling work content in a job shop environment

Abstract Experts estimate that 65% of all manufacturing firms employ fifty or less employees, and many of these small manufacturing firms operate as job shop environments. This paper will focus on a methodology for the development of an expert systems approach to job shop scheduling. Specifically, the concept of prototyping and life cycle development will be discussed. Prototyping combines the steps of knowledge acquisition, knowledge representation, knowledge implementation, and verification and validation into a repetitive cycle, rather than having the steps in a sequential fashion. By using a prototyping cycle, a small expert system is developed first and then gradually enlarged as exception cases are identified, instead of attempting to complete each step entirely before continuing with the next.

Cognitive Radio Network as Wireless Sensor Network (III): Passive target intrusion detection and experimental demonstration

A Cognitive Radio Network (CRN) based Wireless Sensor Network (WSN), as an extension of CRN, is explored for radio frequency (RF) passive target intrusion detection. Compared to a cheap WSN, the CRN based WSN is expected to deliver better results due to its strong communication functions and powerful computing ability. Issues addressed in this paper include experimental architecture, waveform design, and machine learning algorithm for classification. In particular, passive target intrusion is experimentally demonstrated using multiple WARP platforms that serve as the cognitive/sensor nodes. In contrast to traditional localization methods relying on radio propagation properties, the technique used in this research is based on machine learning with measured data, considering complicated multipath environment and high dimensional sensing data collected by the CRN based WSN. Preliminary experimental results are quite encouraging, suggesting that a large-scale CRN based WSN supported by machine learning techniques has promising potential for passive target intrusion detection in harsh RF environments.

Self-improving process control for molecular beam epitaxy

This paper addresses manufacturing research involving advances in material process control. The research objective has been to develop intelligent, self-directed and self-improving control systems which enablein situ (real-time) control path generation based on both product (material behaviour) and processing (control agent) feedback. A ‘product-process’ control philosophy which emphasises product quality is described together with a generic architecture for representing product and process knowledge.Specific details are presented involving the development and application of a self-directed and self-improving material processing system for molecular beam epitaxy of gallium arsenide wafers. Special emphasis is given to the development of a neural model for self-improving control as well as future research directions.

General Mobility Identification and Rectification of Watt Six-Bar Linkages

Mobility identification is a common problem encountered in linkage analysis and synthesis. Mobility of linkages refers to the problems concerning branch defect, full rotatability, singularities, and order of motion. By introducing the concept of stretch rotation, the paper shows the existence of a hidden five-bar loop in a Watt six-bar linkage and how it affects the formation of branches, sub-branches, as well as the whole mobility of the entire linkage. The paper presents the first methodology for a fully automated computer-aided complete mobility analysis of Watt six-bar linkages.Copyright

Artificial Neural Networks In Manufacturing

The topic of artificial neural networks (ANN) has received a great deal of attention among various groups of scientists and engineers as an alternative method of solving either intractable or “fuzzy” problems. In the area of manufacturing, ANN have been applied to vision systems, robot path planning, pattern recognition of data trends for use in analyzing production systems, and some recent research has been attempting to use ANN for feature recognition to facilitate CAD/CAM. This paper will highlight the scope of recent research accomplishments using ANN in manufacturing. Specific research in the control of a Molecular Beam Epitaxy process for growing thin film materials for semiconductors, and the design of a manufacturing system for Group Technology will be examined in closer detail. To conclude the examination of ANN in manufacturing, future trends and research areas will be discussed.