大学生考试网 让学习变简单
当前位置:首页 >> >>

framework for collaborative design of MEMS

framework for collaborative design of MEMS

INSTITUTE OF PHYSICS PUBLISHING J. Micromech. Microeng. 12 (2002) 512–524


Web-based knowledge-intensive support framework for collaborative design of MEMS
Xuan F Zha1,2 and H Du2
1 Gintic Institute of Manufacturing Technology, 71 Nanyang Drive, Singapore 638075, Singapore 2 School of Mechanical and Production Engineering, Nanyang Technological University, 50 Nanyang Avenue, Singapore 639798, Singapore

E-mail: p1137204@ntu.edu.sg, xfzha@gintic.gov.sg and mhdu@ntu.edu.sg

Received 25 April 2001, in ?nal form 7 February 2002 Published 21 June 2002 Online at stacks.iop.org/JMM/12/512 Abstract Microelectromechanical systems (MEMS) design and manufacturing are inherently multi-physical and multi-disciplinary; no single person is able to perform a full development process for a MEM device or system. In this paper we develop a WWW-based design platform for collaborative design of MEMS. The proposed web-based distributed object modeling and evaluation framework with client-knowledge server architecture, KS-WebDOME, allows multi-users/designers in different locations to participate in the same design process. Under this framework, concurrent integrated MEMS design and simulation models can be built using both local and distributed resources, and the design collaboration can be realized by exchanging services between modules based upon CORBA standard communication protocol. To facilitate the rapid construction of the concurrent integrated design models for MEMS, a prototype design system, Web-MEMS Designer, is implemented through concurrent integration of multiple distributed and cooperative knowledge sources and software. By use of the developed prototype system, MEMS design and simulation can be carried out simultaneously and intelligently in an integrated but open design environment on the web. The case of a microgripper design for micro-robotic assembly systems is provided to illustrate how designers in different teams and organizations may participate and collaborate in MEMS design.

1. Introduction
Microelectromechanical systems (MEMS) is a rapidly expanding ?eld of multi-disciplinary technology which takes advantage of semiconductor fabrication processes to produce micrometer-scale mechanical, ?uidic, electric, optical, and other devices. MEMS devices are often integrated with microelectronic circuits which control their behavior, perform signal processing and computing, and control/activate the behavior of the mechanical structures (Przekwas 1998). With the parallel development of new technologies, new device con?gurations, and new applications for microsensors, microactuators, and micro-systems, there has arisen a growing
0960-1317/02/050512+13$30.00 ? 2002 IOP Publishing Ltd

need for multi-disciplinary CAD support for MEMS. The needs, key issues, and requirements in this arena have been identi?ed, formulated and reviewed (Perterson 1978, Senturia and Harris 1992, Senturia 1998). MEMS CAD shares some common techniques with the conventional CAD, but it is different in many ways. The multi-dimensional, multi-disciplinary, and multi-scale nature of MEMS makes the CAD software very dif?cult to develop (Giridharam et al 2001). Smart product design can be achieved with the aid of concurrent and intelligent concepts to facilitate design tasks. The inherently multi-physical and multi-disciplinary MEMS design problem requires new concurrent intelligent design methodologies and systems involving the integration 512

Printed in the UK

Web-based knowledge-intensive support framework

of modeling, design, analysis and evaluation, and simulation for MEM devices or systems as early as possible in the course of the different life-cycle phases. On the other hand, contemporary MEMS design problems often embody signi?cant levels of complexities that make it unlikely that a single designer can work alone. The continuing growth of knowledge and supporting information and everincreasing complexity of design problems has led to increasing specialization. It has been recognized that further rapid progress in MEMS technology will be dif?cult to accomplish without the full range of multi-level hierarchical design tools ranging from high-?delity device level to system level. Because of the heterogeneous structure of micro-systems, MEMS design and simulation require different grades of abstraction and need the cooperation/collaboration of different disciplines and resources. Wide-area networks and the internet-based WWW allow users/designers to provide remote design servers. MEMS CAD systems running on these design servers can support a large-scale group of users/designers who communicate with the systems over the network. Based on the web protocols (e.g. HTTP), user/designer interfaces can provide access to the remote web-based design servers with appropriate web browsers. Users do not need special hardware or software to consult these services. Thus, multiple users/designers in different locations are able to use the same CAD tool and design a MEM device or system together. With the advent of the internet and WWW, it is expected that one of the focal research areas in the MEMS design community will be on the development of WWW-based design framework/platform for collaborative MEMS design. The aim of this paper is to present a web-based knowledge-intensive development framework to facilitate the rapid construction of concurrent integrated distributed design models for MEMS, and to provide distributed designers with a platform/tool for collaboratively building these models. Speci?cally, in this paper, the issues to be addressed and the objectives to be achieved include the following: (1) to explore a new concurrent intelligent design methodology, involving the integration of modeling, design, analysis and evaluation, and simulation, for MEM devices or systems; (2) to develop a concurrent knowledge-intensive design framework for MEMS design and simulation; and (3) to develop a distributed intelligent platform for MEMS design and simulation using Java and CORBA over the internet and WWW. The structure of the paper is as follows. It begins with an overview and requirements for network-based design tools. Then, the web-based framework for supporting different types of collaborative design activities in a distributed design environment is developed for MEMS design and simulation. A case study of the collaborative microgripper design is to illustrate how designers in different teams and organizations may participate in the design of a microgripper for microrobotic assembly system.

2. Literature review
There have been many research efforts on enabling technologies or infrastructure to assist product designers in

the computer network-centric distributed design environment (Pahng et al 1997, 1998, Wood and Agogino 1996, Toye et al 1993, Petrie et al 1994, Frost and Cutkosky 1996, Hardwick and Spooner 1995, Singh 1995). Some of them are intended to help designers to collaborate or coordinate by sharing product information and manufacturing services through formal or informal interactions (Lewis and Singh 1995, Sobolewski and Erkes 1995, Case and Lu 1996). Others propose formalized frameworks that manage con?icts between design constraints and assist designers in making decisions (Pena-Mora et al 1993, 1995, Salzberg and Watkin 1990, Case and Lu 1996). There are also national-level efforts involving university and industry collaboration to make a variety of engineering services available over the internet (MADEFast 1999, NIIIP 1999, RaDEO 1998). The RaDEO program is concerned with comprehensive information modeling and design tools needed to support the rapid design of electromechanical systems. It supports engineers by improving their ability to explore, generate, track, store, and analyze design alternatives. The SHARE project by Toye et al (1993) supports design teams by allowing them to gather, organize, re-access, and communicate design information over computer networks to establish a shared understanding of the design and development process. While SHARE is primarily directed toward interaction through integrated multimedia communication and groupware tools, the NEXT-LINK project incorporates agents to coordinate design decisions affected by speci?cations and constraints (Petrie et al 1994). A networkcentric design system using interacting agents to integrate manufacturing services available over the network is under development (Frost and Cutkosky 1996). The electronic design notebook (EDN) is an interactive electronic document that maintains the look and feel of an engineering document to provide an integrated user interface for computer programs, design studies, planning documents, and databases (Lewis and Singh 1995). Manufacturing tools and services are encapsulated in the hypertext documents and distributed through servers using HTTP (Sobolewski and Erkes 1995). A computer-based design system developed by Sriram and Logcher (1993) provides a shared workspace where multiple designers work in separate engineering disciplines. In their distributed and integrated environment for computer-aided engineering (DICE) program, an object-oriented database management system with a global control mechanism is utilized to resolve coordination and communication problems. Design rationale provided during the product design process is also used for resolving design con?icts. A design information system proposed by Bliznakov et al (1995, 1996) incorporates a hybrid model for the representation of design information at several levels of formalization and granularity. It is intended to allow designers in a large virtual organization to indicate the status of tasks assigned to each designer or team so that other designers can follow their progress. A central database manages pointers and access methods for product and process information in the distributed environment. Hardwick and Spooner (1995) propose an information infrastructure architecture that enhances collaboration 513

X F Zha and H Du

between the design and manufacturing ?rm. This architecture uses the WWW for information sharing and the STEP standard (Owen 1993) for product modeling. It utilizes the CORBA standard for interoperability between software applications in the virtual enterprise. N-dim is a computer-based collaborative design environment for capturing, organizing, and sharing data (Westerberg et al 1995). It is a base, on which applications can be added for the purpose of history maintenance, access control, and revision management. The primary focus of the environment is on information modeling. The system provides a way for de?ning information types that capture the relations between data or models. Pahng et al (1997, 1998) developed a web-based framework for collaborative design modeling and decision support, based on the distributed object modeling and evaluation (DOME). The DOME framework asserts that multi-disciplinary problems are decomposed into modular sub-problems. Modularity divides overall complexity and distributes knowledge and responsibility amongst designers. It also facilitates the reuse of modeling elements. Thus, DOME allows designers to de?ne mathematical models or modules and integrate or interconnect them to form large system models. In DOME, a multiple attribute decision method is used to capture preferences and evaluate design alternatives from different viewpoints. The above ongoing research efforts pave the way in which a network-centric design environment is able to support product designers and suggest what a computerbased design tool or system should look like in such an environment. However, they do not provide a structured and formalized framework for modeling the characteristics of multi-disciplinary and multi-objective design problems, and none of them are focused on the network-centric, distributed and collaborative design of MEMS. Several worldwide projects are continuing to develop comprehensive MEMS design tools focusing on either device or system level CAD. They are derived either from the existing microelectronic design tools (ECAD/TCAD) or mechanical tools (Gilbert 1998). Such systems are at a boundary between two large CAD industries: electronic design automation (EDA) and mechanical design automation (MDA). Thus, a major task of the MEMS CAD system is to intentionally integrate tools from MDA and EDA. The existing MEMS CAD tools for MEMS design, simulation, and manufacturing are unable to support collaborative MEMS modeling and design activities. MEMS CAD software is being developed by several vendors including Coventor, ANSYS, ISE, and CFD Research Corp. The numerous existing high-quality CAD tools for MEMS such as MEMCAD (now CoventorWare) (Gilbert et al 1993, MEMCAD 2000), IntelliSense (IntelliSense 2001), CAEMEMS (Cary and Zhang 1990), OSYSTER (Koppleman 1989), CyberCAD (Tay 1999) etc, are generally specialized and stand-alone applications. It is very dif?cult to use them for understanding and designing the integrated performance of product systems. Therefore, they are unable to support and coordinate highly distributed and decentralized MEMS modeling and design activities (Zha and Du 2001c). The motivation and vision presented in this paper share some similar themes with Pahng et al (1997, 1998) but 514

emphasize design and simulation modeling, decision making, and search/optimization for MEMS.

3. Web-based knowledge-intensive collaborative framework for network-centric design
Contemporary design process is knowledge-intensive and collaborative. The knowledge-intensive support becomes critical in the design process and is recognized as a key solution toward future competitive advantages in product development. The integrated design requires the skills of many designers and experts, in that each participant creates models and tools to provide information or simulation services to other participants given appropriate input information. It is the goal that the collective network of participants exchanging services forms a concurrent model of the integrated design. Based on the DOME framework (Pahng et al 1997, 1998), a web knowledge server framework, KS-WebDOME, is proposed in this paper for collaborative design process (Zha and Du 1999, 2000, 2001a, 2000b). The proposed knowledgeintensive framework adopts the design-with-objects (Zha and Du 1999, 2001c), module network (Pahng et al 1997, 1998, Bic et al 1995), and knowledge server paradigms (Eriksson 1996). The knowledge server paradigms are techniques by which knowledge-based systems can utilize the connectivity provided by the internet to increase the size of the user base whilst minimizing distribution and maintenance overheads. The knowledge-intensive system can then exploit the modularity of knowledge-based systems, in that the inference engine and knowledge bases are located on a server computer and the user interface is exported on demand to client computers via network connections (e.g. internet, WWW). Thus, design modules or objects are connected together so that they can exchange services to form large integrated system models. The module structure leads itself to a client (browser)/knowledge server-oriented architecture using distributed object technology. The main system components of the proposed client/server architecture are shown in ?gure 1. Each of these components interacts with one another using a communication protocol (CORBA) over the internet so that it is not required to maintain the elements on a single machine. As a gateway for providing services, the interface of a system component invokes the necessary actions to provide requested services. To request a service, a system component must have an interface pointer to the desired interface. With a client/knowledge server architecture, the characteristics of the knowledge-intensive framework may be described as follows: (i) extensive knowledge based; (ii) hybrid intelligent system that integrates various knowledge sources; and (iii) concurrent, integrated, and distributed. Thus, an intensive knowledge model for design is a large-scale knowledge framework that allows processing of various types, different levels, and multiple functionality of knowledge in a design process (Zha and Du 1999).

4. KS-WebDOME framework for network-centric design of MEMS
The development process of MEMS devices or systems includes many steps from design, analysis and simulation,

Web-based knowledge-intensive support framework

End User

User-Interface Front End

Network Connection

Problem Solver Knowledge Base Database Knowledge Server


User/Designer DOME Server Interface WWW Knowledge Server Model Base Interface Model Base



Internet DOME Server Database Server

Figure 1. (a) Client-knowledge server architecture. (b) Main system components for KS-WebDOME.

fabrication, to assembly/package, and operation; each of which is dif?cult, high-cost, and time-consuming. Many loops of these steps are needed in the development process. In this section, we will discuss how the KS-WebDOME framework proposed above is used for network-centric MEMS design, analysis, and simulation process. 4.1. MEMS design process and environment A wide range of design problems are included in MEM devices or systems development, such as conceptual design, con?guration design, process simulation, solid-body geometric renderings from photo-masks and process descriptions, optimization of geometry and process sequence, micro-assembly design, planning and simulation, and design of full systems. There are generally two rather different types of CAD requirements (Senturia 1998): conceptual design phase and product-level phase. The ?rst conceptual phase of a new device is to assist in ?nding practical con?gurations; the second product-level phase is to enable careful attention to physical behavior and parasitic phenomena. There is a great bene?t if the actual device masks and process description can be used as input to the simulations. The rendering of three-dimensional (3D) solid models from mask and process data, both to permit checking of geometries and as an input to physical simulation, ensures that the device being simulated is also the one being built. MEMS CAD can be categorized into the work at the following levels: system, device, physical/behavioral, and process level, in which lumped networks, energy-based macromodels, 3D simulation, and TCAD are included respectively (Senturia 1998, MEMCAD 2000, Wilson et al 1999, 2000).

The host of modeling and simulation requirements for a MEMS CAD system at these levels can be identi?ed and described as follows: (1) process modeling tools for all process steps; (2) process optimization tools to achieve a desired device geometry (e.g. topology optimization); (3) physical simulation in multiple coupled energy domain; (4) construction of designer-useful behavioral models from simulation (micro-models); (5) device optimization tools to achieve desired device behavior; (6) insertion of behavioral device models into system-level simulation tools; (7) behavioral model optimization for desired system performance. In an ideal MEMS design environment, the user will ?rst simulate the fabrication process steps to generate the 3D geometrical model including fabrication-dependent material properties and initial conditions (e.g. fabrication-induced stresses). The input to this simulation step is the mask layouts (e.g. in CIF or GDS II format) and a process description ?le (e.g. PFR). To compute fabrication-dependent initial ?elds, the initial geometry model will be meshed and physicsbased process models (deposition, etching, milling, bonding, annealing, . . .) will create a simulation-ready virtual model with complete de?nition of material properties, boundary and volume conditions, and physical/numerical parameters for ?eld solvers. All model parameters should be speci?ed directly ‘on geometry’ rather than on mesh to allow multiresolution (grid independence) and solution-based mesh 515

X F Zha and H Du

Device-level Foundry Libraries and Process Rules Electrical and Mechanical Design Automation Interfaces

Package Libraries

System-level Models & Behavorial Libraries

MEMS 2D Layout & 3D Modeling

System Modeling & Simulation

Other System Domain Simulators

Physical Modeling

Behavorial Modeling

Physical Domain Device Analysis

Electro-mechanical RF Optics Fluidics MEMS Packaging

Structural Analysis

Figure 2. MEMS design methodology and modeling levels (Ref: MEMCAD 2000).

Electro-static analysis Designer (Mask layout and construction, including geometric modeling,drawing and layout, DFMA) User/Designer

Mechatronics analysis

Modeling and performance analyzer and optimizer

Mechanics analysis

Simulator Java Applet Java
Jess/ FuzzyJess

Package MEMS user-interface front end (Netscape or IE) Client Applet TCP/IP character stream Problem solver Fabrication process planner (Fabrication sequence)

Knowledge Server

Figure 3. KS-WebDOME architecture for MEMS design.

adaptation (Przekwas 1998). The ultimate goal, of course, is that the device and the associated system are fabricated, and the system performance is as desired. To the extent that the issues can be anticipated through simulation and modeling, also called computational prototyping, costly fabrication experiments can be reduced in number and increased in effectiveness. Figure 2 shows MEMS design methodology and modeling levels. 4.2. Web-based collaborative design platform for MEMS Based on the design process of MEMS, the KS-WebDOME architecture for the distributed collaborative MEMS design can be illustrated as shown in ?gure 3. Under this framework, the requirements for a web-based MEMS development tool can range from complex intelligent design, modeling, and simulation capabilities to more narrowly de?ned requirements. 516

Its capabilities should be built into selectable or con?gurable, and knowledge-intensive modules that are packaged together to meet the requirements of a desired development ?ow. The web-based collaborative MEMS design platform should be able to address the following issues: device layout and construction; device modeling and simulation; system modeling and simulation, and package, etc. The device layout and construction suite includes a direct, automatic connection between design of process and layout and full 3D-device modeling and visualization. It enables MEMS design to be driven by either experienced layout designers or mechanical engineers demanding full 3D editing capabilities. The device modeling and simulation suite provides solvers for the speci?c 3D physics of each kind of MEMS device. Speci?c knowledge on MEMS device modeling will be wrapped around state-ofthe-art hybrid ?nite element and boundary element numerical

Web-based knowledge-intensive support framework

tools. Thus, MEMS designers do not have to be experts in numerical techniques to get usable, accurate simulation results. The system modeling and simulation suite provides tools to help the designer understand manufacturing sensitivities and co-design of MEMS systems and devices. Design engineers can build and simulate accurate system models containing MEMS components integrated with external or on-chip circuit systems. Advanced tools enable automatic extraction of ef?cient, physically realistic SPICE and SPICElike models of MEMS components from 3D analysis. The packaging suite provides MEMS designers and packaging groups with tools to support communication and co-design. It enables truly coupled 3D package and device co-simulation. Package and MEMS groups can communicate by sharing quantitative models of package-induced effects, along with tools to understand detailed device sensitivities to package design variables. The solution to providing distributed MEMS design support in this research is to extend an original stand-alone MEMS design system, i.e. MEMS Designer (Zha and Du 1999), into a web-based collaborative MEMS design system, i.e. Web-MEMS Designer. The system is deployed on a web server enabling access via the internet to a comprehensive suite of scalable and con?gurable software tools for MEMS design and simulation. Details about the implementation of the webbased MEMS Designer system will be discussed below.

The developed prototype Web-MEMS Designer system contains a set of modules that are able to preliminarily support some of these functions, as follows (Zha and Du 1999, 2000, 2001a, 2000b, 2000c, Bay et al 2000, Toh et al 2001, Lin 1997): (1) (2) (3) (4) (5) (6) (7) (8) function–behavior–structure modeler; 2D drawing tool (including mask layout editor); fabrication process sequence builder; manufacturing and material database; manufacturing advisory system; design optimization tool (e.g. GA tools); 3D geometric modeler and viewer; 2D and 3D FEM mesh generator (including an ANSYS interface) (in progress).

The capabilities of these modules enable the Web-MEMS Designer system to offer a special design platform for collaborative MEMS design and simulation. 5.2. System implementation The implementation of the prototype Web-MEMS Designer system is actually a three-stage process. The ?rst stage is to convert the MEMS Designer (Zha and Du 1999) into a stand-alone application, involving the translation of the original knowledge base into an appropriate format and reconstructing the necessary functionality. Then, the second stage is to convert the stand-alone application implemented in C/C++ into CGI executables that were deployed on a standard web server, in terms of template web pages to contain dynamically generated input forms, the necessary code to extract knowledge from submitted forms, and display results. The third stage is to implement the Web-MEMS Designer using Java and CORBA technologies integrating with a Java Expert System Shell, Jess/FuzzyJess (Jess 1999), based upon a Windows NT-based environment with a front-end web-browser-based graphical user interface (GUI). Jess is a multi-paradigm programming language that provides support for rule-based, object-oriented, and procedural programming system language. The underlying modules are written in Java/Java3D/JDBC, respectively. The implementation architecture shown in ?gure 4 uses the two-tier client/knowledge server architecture of ?gure 1 to support collaborative design interactions. Designers can integrate MEMS design problem models with the existing application packages, such as Java3D and JDBC for CAD and database applications. The CORBA (Siegel 1996) standard is used to add distributed communications capabilities to modules (Orbix and OrbixWeb from IONA Technologies Ltd (IONA 1997)). CORBA serves as an information and service exchange infrastructure above the computer network layer and provides the capability to interact with the existing CAD applications and database management systems through other object request brokers (ORB). In turn, the KS-WebDOME framework provides the methods and interfaces needed for the interaction with other modules in the networked environment. These interactions are graphically depicted in ?gure 5. When a change is made by designer B, the service corresponding to the request from designer A will re?ect the design change. 517

5. Development of web-based collaborative MEMS design system
To facilitate the rapid construction of the concurrent integrated models, a web-based design environment is essential for MEMS design and simulation. In this section, we will describe a prototype implementation of Web-MEMS Designer system. The focus is on the description of the technologies employed in the design and development of the WebMEMS Designer system under the KS-WebDOME framework discussed above. 5.1. System overview The Web-MEMS Designer system is a knowledgedriven design platform that delivers complete end-to-end development ?ow for MEMS-enabled devices or systems. It equips design engineers with the means to develop MEMS devices or systems from an initial concept through complete coupled analysis, which can also include package design characteristics, and ultimately extract high-level models for system simulation. The Web-MEMS Designer system exploits the modularity of knowledge-based systems, in that the inference engine and knowledge bases are located on server computers and the user interfaces are exported on demand to client computers via the web. It is therefore a distributed intelligent development environment, consisting of 3D design, modeling, and simulation software tools, which enable the creation of complex micro and/or MEM devices. The design ?ow of the Web-MEMS Designer system is similar to MEMCAD (MEMCAD 2000) that coordinates four key MEMS product development functions: layout and construction, device modeling, systems modeling, and packaging analysis.

X F Zha and H Du
Client Knowledge Server DOME Server WebDOME GUI Optimization (GA) Knowledge MDL Base Interpreter Inference Engine OMG Kernel DOME Server Interface Java Applet-based Object Request Broker CORBA-Compliant Object Request Broker CORBA-Compliant Object Request Broker CAD Server CAD Applications CORBA-Compliant Object Request Broker Database Server DBMS Graphics Server Graphics (2D & 3D) CORBA-Compliant Object Request Broker Model Base Server Model Base Server Interface Java-based Object Request Broker

WWW-based GUI

Network Backbone (Internet/Intranet,WWW)

Figure 4. Implementation of the open design environment based on KS-WebDOME framework.
Service request Service provision Change requested

Designer A

Designer B

Modeling Layer (GUI)

Workspace in WebDOME Servers

CORBA Communication Layer (Internet,WWW)

Figure 5. Service exchanges between distributed modules.

The enumerated request shows the sequence for obtaining the service that is needed by designer A. The light gray module seen by designer A is a remote module published by designer B. The underlying collaboration mechanism is based on the board systems. Each modular system has both blackboard and whiteboard systems. The blackboard system is used for the local modular system to store intermediate reasoning and calculation results, which dynamically ?ushes in running. The whiteboard system is actually a bulletin board for information and collaboration. The Web-MEMS Designer GUI provides users with the ability to examine the con?guration of design problem models, analyze trade-offs by modifying design parameters within modules, and to search for alternatives using an optimization tool. The GUI is a pure client of the WebDOME server, delegating all events to an associated WebDOME server. For wide accessibility and interoperability, the GUI is implemented as a web browser-based client application, which is a combination of HTML/XML documents and Java applets. For the CORBA-based remote communication between the GUI Java applets and the back-end side system components 518

such as DOME server, CAD server, graphics server, and model base server, a commercial ORB implementation of Java applets (OrbixWeb) is employed (IONA 1997). Based on the system implementation architecture in ?gure 4, the functionality of the knowledge server is achieved through implementing DOME servers, model base server, core knowledge engine, database server, and even knowledge base assistant and inter-server communications explanation facilities. Figure 6 shows a demonstration screenshot of Web-MEMS Design GUI applet. The GUI interacts with designers’ events and requests to the DOME server that provides the back-end implementation for the modeling of design problems. The core of the server is based upon object modeling and evaluation (OME) kernel (Pahng et al 1997, 1998) written in Java/Java3D/JDBC, integrating Jess/FuzzyJess. The back-end implementation for knowledge server, including DOME server, and model base server, and the front-end interface to the GUI are written in Java. The DOME server manages each design session in a workspace and can simultaneously maintain several workspaces. The workspace manages administrative aspects

Web-based knowledge-intensive support framework

Figure 6. Web-MEMS Designer GUI Applet (Demo).

of a model (e.g. ownership, access privilege, links to other workspaces in different DOME servers, etc). The DOME server itself is a CORBA-compliant distributed object and can communicate with other DOME servers. The model base server maintains persistent storage for models created by the DOME servers. The model repository stores a model in a model de?nition ?le (MDF) with two parts: meta de?nition and model de?nition. The meta de?nition contains the information such as model id, ownership, and access privilege information. The model de?nition is based upon a model de?nition language (MDL) used by the system. The core knowledge engine includes the knowledge base and a problemsolving paradigm (inference engine). The knowledge base is built in Java/Jess. The web database system is developed by use of Microsoft Access databases to store the details of data and Java programs to access these databases through JDBC connections (Toh et al 2001, Zha and Du 2001a).

6. Case study: collaborative design for a microgripper
To illustrate the application of the developed Web-MEMS Designer system for collaborative MEMS design process, a working case of a microgripper design for a microrobotic assembly system was carried out. The design case

originated from Su (1999). It was chosen because of its interdisciplinary and developing nature. The research results from this particular case could be generalized to cover other designs that require collaboration and integration of multiple domains. The focus of the illustration is on how designers from different teams, divisions, or companies in remote locations may participate to create an integrated design model for the microgripper design. Generally, a micro-robotic assembly system consists of three major parts: a micro-robot system, an assembling platform, and micro-components to be assembled. A micro-robot system is generally composed of a micro-robot body and its end-effector with a microgripper. Thus, the overall topology of the design problem and the design workspace can be illustrated in ?gure 7. As shown in ?gure 7, the micro-robot and microgripper manufacturers provide their design and simulation models to the microrobot system designers who in turn develop the technical models for the micro-robot system. The micro-robot system manager collaborates with the micro-robot system designers and provides models and data for micro-robot operating conditions and requirements. Then, he/she uses the microrobot system design models created by the micro-robot system designers to develop cost evaluation and redesign models. The microgripper and micro-robot manufacturers develop models for their products so that their customers can 519

X F Zha and H Du
Microgripper Manufacturer Microrobot Manufacturer

Microgripper Design

Microgripper Simulation

Microrobot Design

Microrobot Simulation

Network (Internet, WWW)

Microrobot System

Microrobot System Design Microrobot System Manager Microgripper


Cost Model

Microrobot System


Work Platform

Redesign Model

Microrobot System Operation

Microgripper Parameter

Microrobot Parameter Microassemblies

Figure 7. Problem topology of the micro-robot system design model.

obtain performance predictions and evaluations for different parametric con?gurations and operating conditions. These individual models are constructed, published, and served by different companies, as shown in ?gure 8(a). If a single designer or company creates all these models and provides all those services the design work will be carried out in an individual workspace, as illustrated in ?gure 9. The design session GUI of Web-MEMS Designer creates and depicts the layout and construction and simulation models or modules (Su 1999) in the microgripper design workspaces. Designers can use any commercial Web browser to access and work on these modules. Since users/customers will connect to these models to assess the performance of their products, designers will decide how to publish these models, i.e. what simulation services will the models offer given appropriate input information. When a model is published anyone can use its services if he/she has the appropriate access privileges. The owner of the model can or may want to conceal knowledgeintensive engineering formulae or supply chain information embedded in the model. Through service publication, a designer sets access privilege levels for the services of each module in their workspace. Therefore, the designer working on the design model is assigning access privileges to the services that modules can provide. Since the micro-robot system design and operations are tightly coupled, it would make sense for designers in these groups to share a common model. Therefore, while designers from different groups are in remote locations, they can access into the same workspace, which is referred to as a shared workspace. Figure 10 shows the design workspace as viewed by the micro-robot system designers 520

and operation designers. The micro-robot system designer is connected to the microgripper and micro-robot manufacturers. The micro-robot system designers can test their micro-robot system design integrated with microgripper and micro-robot models. The designer from micro-robot system user or operation team shares the workspace with the micro-robot system designer. The micro-robot system design team creates modules in the upper left corner while the micro-robot system operation team makes the rest designer. In this case the microrobot system design team owns the session and the operation team has joined as a builder. Although builders cannot modify the modules created by other builders or owners, they can add new modules and utilize all services. For example, the micro-robot system operation team designer can use a service from a micro-robot system design groups’ module to obtain the micro-robot accuracy and the open distance of the microgripper and can build new modules in the workspace that utilizes this information. The micro-robot system designer can use services from the models published by the micro-robot and its microgripper manufacturers. Utilizing models provided by other designers is referred to as subscribing to a model. It is the responsibility of the micro-robot system designer to provide these data or to locate other models that can provide these data as services. The micro-robot system manager wants to evaluate the micro-robot system design from a cost viewpoint. They link their models to the micro-robot system design to obtain the information services needed by their models (?gure 8(b)). The micro-robot system designers have only published cost-related aspects of their model. Therefore, the micro-robot system manager can only observe elements of the micro-robot system design

Web-based knowledge-intensive support framework

Microgripper Manufacturer Host Workspace Interface Model Base Interface Workspace Interface Model Base

Microrobot Manufacturer Host Model Base Interface

Microgripper Workspace

Microrobot Workspace

Model Base

DOME Server

Model Base Server Knowledge Server

DOME Server Knowledge Server

Model Base Server

Microrobot System Design Team Host Microrobot System Designer Microgripper Microrobot Workspace Interface Microrobot System Host Model Base Interface

Microrobot System Operator Host Microrobot System Designer Microgripper Microrobot

Microrobot System Workspace MicroMicrogripper robot

Model Base

DOME Server Knowledge Server

Model Base Server

Microrobot System Manager Host Workspace Interface Model Base Interface

Microrobot System Manager
Micro-robot System Microgripper Microrobot

Microrobot Manager Workspace Micro-robot System

Model Base

DOME Server Knowledge Server

Model Base Server


Figure 8. (a) The design workspace as viewed by the micro-robot system designer and operator. (b) Micro-robot system manager model connected to the micro-robot system design model.

Microgripper Designer Web-MEMS Designer GUI Java-enabled Web Browser (Netscape or IE) Client Applet

Workspace Interface Design Workspace WWW Internet DOME Server

Model Base Interface Design Model Base

Design Model Base Server Knowledge Server

Figure 9. Individual workspace for a microgripper designer.

model that were published, as the micro-robot designer wanted to protect their proprietary models. Figure 10 shows the microgripper design and simulation models for layout, 2D, and elastic simulation. Figure 11 shows a 2D mesh generation GUI of the Web-MEMS Designer for the microgripper. The microgripper analysis and simulation and the microgripper design are also tightly coupled so that the designers from different design and simulation groups may also need to share a common model and access the same workspace, although these groups may be in remote locations.

The micro-assembly system is operated by means of a virtual micro-robot manipulation system in which 3D models of the micro-components are manipulated virtually in a computer graphics constructed in the WWW scheme. The microassembly system simulator provides a new design tool of 3D MEMS by combining the possibility of ?exible assembly and intuitive operations. When a simulating assembly or operating sequence is running, users can control the microgripper open–close states, micro-robot positions and orientations, and micro-components positions and orientations by clicking 521

X F Zha and H Du

(a) Layout and elastic simulation model

(b) Mask and generated 3D model
Figure 10. Microgripper design and simulation: layout, 3D model and elastic simulation model.

on them. The user interface graphically displays microrobot con?gurations, microgripper states, and the component states.

7. Summary and future work
In this paper we have presented a web-based design platform for supporting collaborative MEMS design over the internet and WWW. A two-tiered client (browser)/knowledge server architecture was adopted to allow experts and designers to publish and subscribe to modeling services on the WWW. The proposed KS-WebDOME framework is built to provide module network architecture for integrating modeling services. In the module network, design resources, models, data, and activities are not centralized nor concentrated in one location. They are distributed among many companies, designers, or design participants working together over the 522

internet/intranet. When module services are connected, the resultant service exchange network creates a concurrent integrated system model or a module network that invokes a chain of service requests if needed to provide correct information. To provide distributed designers with a tool for collaboratively building the concurrent integrated design system models, the KS-WebDOME framework is extended to be a computer network environment focusing on the design and simulation for MEMS. MEMS design modules are created by fully implementing locally de?ned modules and subscribing to the services of remote modules. The implementation of the Web-MEMS Designer system hides the details of the remote interaction mechanism from the user but allows the MEMS designer to model interactions between local and remote modules in a transparent manner. In turn, designers can selectively publish modeling services for use by others. The microgripper design for the micro-robot assembly example

Web-based knowledge-intensive support framework


Figure 11. Microgripper mesh generation by Web-MEMS Designer (Demo).

illustrates the concept and different models of collaboration supported by the prototype implementation. The preliminary implementation of the Web-MEMS Designer system illustrates the potential of KS-WebDOME framework for MEMS design and simulation. When fully implemented and integrated with other computer-based collaboration tools, the Web-MEMS Designer system will provide designers with a powerful infrastructure for concurrent product design for MEMS. However, there exists a large amount of work to be done both on the particular design paradigms or methodologies for MEMS and the system development. For example, the framework should accommodate top-down and bottom-up approaches or models in the context of both traditional sequential design processes and concurrent design for MEMS devices or systems. In a collaborative design environment, there are also a number of fundamental issues yet to be addressed such as knowledge base evolutionary maintenance, model interface standard, computational strategy for resolving circular dependencies in a WebDOME model, parallel service request invocation, etc. In addition, other aspects such as human interaction and knowledge sharing will still require the integration of additional support tools with the framework (e.g. e-mail, video conferencing, hypertext documentation with XML, etc). The project described in this paper is in progress.

The authors would like to thank the anonymous reviewers of this paper for their insightful comments and suggestions that helped in improving the paper.

Bay E, Seah C, Zha X F and Du H 2000 A web-enabled drawing applet for masking process in MEMS Final Year Project Report School of Mechanical and Production Engineering, Nanyang Technological University, Singapore Bic L F, Fukuda M and Dillencourt M B 1995 Distributed computing using autonomous objects IEEE Computer 29 55–61 Bliznakov P I 1996 Design information framework to support engineering design process Dissertation Arizona State University Bliznakov P I, Shah J J, Jeon D K and Urban S D 1995 Design information system infrastructure to collaborative design in a large organization Proc. ASME DETC (Boston, MA) vol 1 pp 1–8 Cary S B and Zhang Y 1990 CAEMEMS: an integrated computer aided engineering work bench for micro-electro-mechanical systems Proc. IEEE Micro-Electro-Mechanical Systems Workshop Case M P and Lu S C-Y 1996 Discourse model for collaborative design Comput. Aided Des. 28 333–45


X F Zha and H Du

Eriksson H 1996 Expert systems as knowledge servers IEEE Expert 14 14–9 Frost H R and Cutkosky M R 1996 Design for manufacturability via agent interaction Proc. 1996 ASME DETC (Irvine, CA) Gilbert J R, Osterberg P M, Harris R M, Ouma D O, Cai X, Pfajfer A, White J and Senturia S D 1993 Implementation of a MEMCAD system for electrostatic and mechanical analysis of complex structures from mask descriptions Proc. MEMS’93 (Fort Lauderdale, FL, Feb. 1993) pp 207–12 Gilbert J R 1998 Integrating CAD tools for MEMS design Computer 31 99–101 Giridharam M G et al 2001 Multi-disciplinary CAD system for MEMS J. Modeling and Simulation of Microsystems 2 43–50 Graham I 1994 Object-Oriented Methods (Reading, MA: Addison-Wesley) Hardwick M and Spooner D 1995 An information infrastructure for a virtual manufacturing enterprise Proc. Concurrent Engineering: A Global Perspective (McLean, VA) pp 417–29 IONA 1997 Orbix2 Programming Guide (IONA Technologies Ltd) IntelliSense 2001 webpage http://www.intellisense.com Jess 1999 Sandia National Laboratories CA webpage http://herzberg.ca.sandia.gov/jess Karam J M et al 1997 CAD and Foundries for Microsystems DAC97 (Anaheim, CA) Koppleman G M 1989 OYSTER, a three dimensional structural simulator for micro-electro-mechanical design Sensors Actuators 20 179–85 Lewis J W and Singh K J 1995 Electronic design notebooks (EDN): technical issues Proc. Concurrent Engineering: A Global Perspective (McLean, VA) pp 431–6 Lin H C 1997 JAVAMESH—A two dimensional triangular mesh generator for ?nite elements Master’s Thesis University of Pittsburgh MADEFast 1999 webpage http://madefast.stanford.edu MEMCAD 2000 webpage http://www.memcad.com NIIIP 1999 webpage http://www.niiip.org Owen J 1993 STEP—An Introduction (Winchester: Information Geometers Ltd) Osterberg P M and Senturia S D 1995 MEMBUILDER: an automated 3D solid model construction program for micro-electro-mechanical structures Proc. Transducers’95 (Stockholm, Sweden) pp 21–4 Pahng F, Bae S H and Wallace D 1998 Web-based collaborative design modeling and decision support Proc. DETC’98 (Atlanta, GA) Pahng F, Senin N and Wallace D 1997 Modeling an evaluation of product design problems in a distributed design environment, CD ROM Proc. ASME DETC (Sacramento, CA) Pena-Mora F, Sriram R and Logcher R 1993 SHARED DRIMS: SHARED Design Recommendation-Intent Management System, Enabling Technologies: Infrastructure for Collaborative Enterprises (Piscataway, NJ: IEEE) pp 213–21 Pena-Mora F, Sriram R and Logcher R 1995 Con?ict mitigation system for collaborative engineering AI EDAM—Special Issue of Concurrent Engineering 9 101–23 Perterson K E 1978 Dynamics micromechanics on silicon: techniques and devices IEEE Trans. Electron Devices 25 1241–50 Petrie C, Cutkosky M and Park H 1994 design space navigation as a collaborative aid Proc. 3rd Int. Conf. Arti?cial Intelligence in Design (Lausanne, Switzerland) Przekwas Abdrzej 1998 An Integrated Multidisciplinary CAD/CAE Environment for Micro-Electro-Mechanical System (MEMS) (CFD Research Corp., AL) RaDEO 1998 webpage http://elib.cme.nist.gov/radeo

Salzberg S and Watkin M 1990 Managing information for concurrent engineering: challenges and barriers Res. Eng. Design 2 35–52 Su C 1999 Development of three MEMS devices: a microgripper, a micromechanism and a microaccelerometer M. Eng. Thesis Nanyang Technological University, Singapore Senturia S D 1998 CAD Challenges for microsensors, microactuators, and microsystems Proc. IEEE 86 1611–26 Senturia S D and Harris R M 1992 A computer-aided design system for micro-electro-mechanical systems (MEMCAD) J. Microelectromech. Syst. 1 3–13 Siegel J 1996 CORBA: Fundamentals and Programming (OMG) (New York: Wiley) Singh A K 1995 CONSENS—an IT solution for concurrent engineering Proc. Concurrent Engineering: A Global Perspective (McLean, VA) pp 635–44 Sobolewski M W and Erkes J 1995 CAMnet architecture and applications Proc. Concurrent Engineering: A Global Perspective (McLean, VA) pp 627–34 Sriram D and Logcher R 1993 The MITDICE Project IEEE Computer 27 64–5 Tay Francis E H 1999 Collaborative design software for MEMS Proc. SPIE 3894 152–8 Toh W Y, Zha X F and Du H 2001 Development of a web database system for manufacturing processes of MEMS Final Year Project Report School of Mechanical and Production Engineering, Nanyang Technological University, Singapore Toye G, Cutkosky M R, Tenenbaum J M and Glicksman J 1993 SHARE: a methodology and environment for collaborative product development Proc. 2nd Workshop on Enabling Technologies: Infrastructure for Collaborative Enterprises (Morgantown, WV) pp 33–47 Westerberg A W et al 1995 Distributed and collaborative computer-aided environment in process engineering design Proc. Int. Conf. on Intelligent Systems in Process Engineering (ISPE’95) (9–14 July, Snowmass Village, Colorado, USA) Wilson N M, Liang S, Pinsky P M and Dutton R W 1999 A novel method to utilize existing tcad tools to build accurate geometry required for MEMS simulation Proc. Modeling and Simulation of Microsystems (19–21 April 1999) pp 120–3 Wilson N M, Wang K, Yergeau D and Dutton R W 2000 GEODESIC: a new and extensible geometry tool and framework with application to MEMS Proc. Modeling and Simulation of Microsystems (27–29 March 2000) pp 716–9 Wood III W H and Agogino A M 1996 Case-based conceptual design Information server for concurrent engineering Comput. Aided Des. 8 361–9 Zha X F and Du H 1999 Knowledge intensive methodology for design and simulation of micro-electro-mechanical system (MEMS) devices Report-MEMS CAD/CAE/CE for Singapore NSTB Project ARC 5/97, School of Mechanical and Production Engineering, Nanyang Technological University, Singapore Zha X F and Du H 2000 Web-based knowledge intensive collaborative design modeling and decision support Proc. Int. Conf. Engineering and Technological Sciences (Beijing, China) Zha X F and Du H 2001a A world wide web based manufacturing advisory service system for fabrication/packaging process/material selection in concurrent collaborative design of MEMS devices Proc. Int. Conf. Material for Advanced Technologies (Singapore) Zha X F and Du H 2001b Web-Based knowledge intensive collaborative design framework for MEMS Proc. Int. Workshop on MEMS (Singapore) Zha X F and Du H 2001c Mechanical systems and assemblies modeling using knowledge intensive petri net formalisms Artif. Intell. 15 145–71


网站首页 | 网站地图
All rights reserved Powered by 大学生考试网 9299.net