Flexible Interfaces in the Industrial Environment.
By Hall W, Weal M, Heath I, Wills GB, Crowder RM.
International Conference Managing Enterprises- Stakeholders, Engineering, Logistics and Achievement (ME-SELA'97) Loughborough, UK. 22-24 July 1997 pp453-460.
Synopsis
The management of information is one of the greatest problems facing industry today. Whilst many firms have introduced computer systems into the office environment, providing access to electronic information on the shop floor to support the manufacturing process is a much more complex problem. This paper looks at the application of open hypermedia systems to the development of information systems to support the manufacturing process. A particular focus of the work was the development of an interface that was easy to use by engineers on the shop floor whilst still giving them full access to all the maintenance, set-up and fault diagnostic information in the application, and potentially integration with other computing systems in use by the company.
Keywords: User Interfaces; Hypermedia; Information Management; Open Systems; Manufacturing
The Screen Handler Enabling Process - SHEP
The use and management of information within an industrial environment is undergoing change. Traditional paper based documentation is slowly being replaced with electronic equivalents. As the scale of available electronic information increases, managing the information so that the correct people can find the information easily becomes a critical issue. This is not limited to information storage but also to the interface providing the access mechanisms.
Industrial environments typically employ a wide range of people with vastly differing computing skills. For example, the use of computers on the shop floor has been limited, to Computer Numerical Control (CNC) machining and monitoring equipment with very simplistic fixed page displays. More recently, there has been a move towards using a Programmable Logic Controller (PLC) with a user interface displaying only data limited to fixed page screen formats. The users are able to change between pages but are not able to change the way the data is displayed. However, the office environment tends to provide more modern computing facilities for the staff with standard windows-based graphical user interfaces.
Many firms now have a stated goal of seeking to implement a company-wide fully integrated computer system, with a further goal of connecting this to the global Internet or a company intranet (or both) at sometime in the future. With the advent of the World Wide Web (WWW) and the phenomenal growth of the use of the Internet for business and commercial purposes, many firms are considering, or have already taken the plunge into, setting up a company Web site for publicity or marketing purposes, even though their internal computing systems may still be fairly primitive.
The problem therefore is two-fold. Firstly providing people on the factory floor with electronic information systems to help them find the information they want, when they want it, and in such a way that the interface to the system is practical and easy to use in a busy factory environment. And secondly, creating such systems so that they can be fully-integrated with the other computing systems used by the firm, whether they be part of the diagnostics of a machine, the design system, the office system, the company intranet or whatever.
The problems and issues involved with managing large information systems for such an environment are well described by Malcolm et al. (1). One approach is to structure and manage the information using a hypermedia system. However, the non-linear presentation of information adds additional interface considerations. The open hypermedia system, Microcosm, developed by the University of Southampton Multimedia Research Group (2) provides a suitable environment for developing different user interfaces. In this paper we discuss how the Microcosm system has been applied in a case study undertaken with Pirelli Cables UK at their Aberdare plant, and in particular the development of a new interface control mechanism SHEP, that was designed to enable a windows-based interface to be more easily used on the shop floor. We discuss the results of this work and future developments.
The main objective of the initial EPSRC funded project was:
To explore the potential of a model for hypermedia as an operational interface in the Advance Manufacturing Technology (AMT) environment, and develop extensions to the model (3).
To demonstrate that the concepts were viable, a case study was undertaken, based on a cable packaging line at Pirelli Cables, Aberdare. The hypermedia application developed had particular emphasis on maintenance and operator set-up.
The project involved collating the paper manuals, working instructions, engineering drawings, data tables, and other associated information and then transferring them into electronic format. The Aberdare Case Study used the hypermedia system, Microcosm, which was originally developed at the University of Southampton but was commercialised through Multicosm Ltd during the lifetime of the project (4) and is more fully described below. The collated documents were assembled within Microcosm into an information resource database. A three dimensional model of the line was constructed and used as an initial gateway to the underlying information.
The information system was targeted at the following users:
The project was also concerned with how this information was delivered to the relevant personnel (5). This encompassed not only the hardware used (a pen-based portable computer), but also the interface to the information. The delivery system was targeted at shop-floor personnel requiring access to the information at the machine itself (6).
The Microcosm open hypermedia system
Microcosm is an open hypermedia system (7), which separates the links from the data (text documents, video clips, photographs, tables, technical drawings, audio files etc.). This enables the documents to be stored in different locations, and in their original format. The links can be arranged in link databases (linkbases) to represent different cognitive and pedagogical structures. This approach to hypermedia system design has many advantages over other closed hypermedia systems in terms of authoring efficiency, particularly in large-scale applications, and facilitate maintenance, reusability and customisation of both the information content and the hypermedia structure (8). The open architecture allows different processes to be incorporated into an application. The processes range from simple link creation and search algorithms to dynamic link creation, such as information retrieval and rule-base algorithms.
Figure 1: A typical Microcosm screen layout
The Microcosm system relies on the underlying operating system to provide the user interface. Since the version of Microcosm available for the project was implemented to run under Microsoft Windows, this is the default user interface and the user is required to be familiar with MS Windows in order to access the information in the system. When accessing information using Microcosm it is quite common to have a number of windows open at one time (Figure 1) and if the user isn't familiar with the concept of manipulating windows this can be confusing or just consume too much time unnecessarily. For shop floor use therefore, more control is required over the display of windows as the user moves around the application. So as part of the Aberdare case study, a simple user interface was developed for Microcosm to provide access to the underlying information. This is described below.
The approach used in Aberdare Case Study was to place an interface layer on top of the Microcosm environment. This layer consisted of a toolbar designed in consultation with he shop floor personnel (Figure 2). Its functionality consisted of :-
Figure 2: The Aberdare interface
After construction, an expert review of the application, using discount usability engineering (9) was used to evaluate the user interface. The following section discusses some of the problems encountered.
Lessons learned from the case study
As was expected, the industrial environment brought together users of different computer skills. Therefore, the user’s interface requirements varied. However, two additional requirements of the interface were identified :-
To cater for these requirements a new control strategy of the user interface is required. One such approach is to provide central information control, where the individual interface components defer control over their positioning on the screen, enabling their organisation within some predetermined strategy. The proposed solution was to implement an architecture which promoted the disclosure of state information by the individual components and allowed screen management processes to modify the interface.
The Screen Handler Enabling Process - SHEP
In a windowing system, each interface component (window, dialogue, toolbar etc.) can be said to have a state that defines the look of the component. This state can be composed of external and internal information. The external state describes the size and position of a component on the display. The internal state contains such information as the font used within the window, colour schemes or even the position of scrollbars in the interface. Screen management need not be restricted to the external state of an interface component. Sometimes the position within a document might need to be included in the management of the component state. An example of the usefulness of this would be the automatic scrolling of a text window that contains a commentary about a video clip. This form of synchronisation is next to impossible without communicating the internal state between the components in question.
The idea behind SHEP is to provide a central controlling mechanism that can arbitrate between external screen management processes and the interface components. An interface component within the Microcosm system might be a document viewer, a toolbar, or a dialogue box (such as a list of links). The core of the SHEP architecture is implemented as a central process through which all other components communicate. This avoids the need for components to have knowledge of one another and enforces a common protocol. This architecture is open and thus easily expanded.
The components in SHEP are divided into two groups:
Figure 3: The SHEP Architecture
The sheep communicate with the SHEP controller before carrying out any operation that involves a change of state. This allows shepherds the opportunity to modify the state information or move other sheep to accommodate the changes better. The sheep also provide SHEP with a means to communicate with them. This creates a two-way communication channel allowing the sheep to supply state changing information and SHEP to give the sheep orders to carry out.
The shepherds must register themselves with SHEP when they are invoked. They are placed in a chain through which the state messages are passed. This architecture allows modules to be slotted in and out during run-time, if required, and keeps the overall system flexible and open.
The SHEP system also defines a protocol that ensures the sheep and the external shepherds can all communicate. This protocol details a series of actions, which the sheep can send or receive, and a list of common attribute/value pairs that the components can use to exchange information. This does not preclude the inclusion of attributes that are specific to a certain interface, for example a volume attribute might be included with the state for an audio viewer. The majority of shepherds would not understand this attribute and therefore ignore it. However, specific shepherds might be written to manage the volume state of audio viewers.
When the sheep register, they can also supply information about themselves. For example, their ability to respond to changes in size, position etc. This information might be useful to shepherds when positioning windows on the screen. Standard sheep information is defined within the SHEP protocol.
A final part of the SHEP protocol concerns the displaying of new information. In order to manage the screen properly the external process need to be able to control how many documents are on the screen at the same time. To enable this the document dispatcher registers with SHEP and notifies the shepherds whenever it is about to launch a new document. As part of the notification process, the shepherds can block the creation of the new document, ask the dispatcher to re-use a current viewer or perhaps close down open documents to make room for it.
The application of SHEP in the industrial domain
Complex machinery is beginning to be supplied with electronic manuals to aid faultfinding (10). However, the information is limited to a narrow data set and cannot be easily integrated with other factory information systems. The Factory Information Resource Management (FIRM) project is examining some of the issues involved in making the provision of information factory wide, drawing from any other information systems available (11).
The system will be used by a wide range of users from the factory environment. Thus SHEP will provide an adaptive user interface that can be configured on a person-by-person basis. The shop-floor operators will still have the simple single-document-plus-toolbar interface. Maintainers can have more control, and access to other tools. Office personnel can have information from other information systems appearing as a part of their current windowing environment with, perhaps, little overall control.
SHEP also provides the user with more control, if so required. Once a particular user becomes more familiar with the environment, they can change their interface so that more features are accessible. There is much finer granularity of control over all aspects of the user interface enabling diverse factors such as user experience, task, or operating environment to influence the interface to the system dynamically.
Flexible interfaces are essential in an industrial environment. With the vast diversity of system users, a single interface will never be able to encompass all of the users individual needs. However, it is virtually impossible to foresee all of the requirements for the interface for a project and so any proposed solutions must be configurable and open. The SHEP process is an example of such an approach. It provides an open protocol through which the various components of the interface can communicate to achieve a particular end. By defining an external process to perform the control, the system is flexible, allowing new approaches to be designed and integrated at a later date, without needing to alter any other part of the environment. Finally, the fact that the shepherds communicate dynamically means the system can adapt to changes in the environment, users needs etc. while it is being used.
SHEP is only part of the answer of course. It provides a flexible way of controlling the display of multiple windows on a computer screen. But there are still many issues concerned with how the user interacts with the information in the windows. Microcosm provides a powerful query interface for both accessing and creating hypermedia links, in which the user can initiate a query to the linkbases or other filter processes. However, this requires the user to select an object on the screen and then select an item from a menu to form the query. This two stage process is not practical to use on the shop floor and so for the Pirelli application, whilst extensive use was made of the dynamic link creation facility in Microcosm for authoring purposes, all information is presented to the user in the form of a button interface to the links.
As mentioned previously, for the user trials we used a portable pen-based computer, so that the operators and engineers could both carry the machine around with them and didn't have to worry about a key-board and mouse. The user follows a link by pressing on the button using the pen. But this does limit the interaction that the user can have with the system. We are now experimenting with speech interfaces, which could potentially enable the combination of both the simple direct access interface to links and the more powerful query interface, but the high-level of background noise on the shop floor and the problems of training the current generation of speech recognition systems, make this still rather futuristic. However, combine our work above with a speech interface and a wearable (hands-free) computer and you have the factory information system of the future.
The authors acknowledge the EPSRC for funding the work under grant numbers GR/H/43038 and GR/L/10482, and Pirelli Cables Aberdare for allowing us to use their information.
(1) Malcom KC, Poltrock SE, Schuler D. Industrial Strength Hypermedia: Requirements for a Large Engineering Enterprise. In Hypertext ‘91. The Third ACM Conference on Hypertext, San Antonio, Texas, December 1991, pp 13-24. ACM Press 1991.
(2) Fountain AM, Hall W, Heath I, Hill G. MICROCOSM: An Open Model for Hypermedia with Dynamic Linking. In Rizk A, Streitz N & Andre J. eds. Hypertext: Concepts, Systems and applications. The Proceedings of the European Conference on Hypertext, INRIA, France, pp 298-311. Camberage University Press 1990.
(3) Crowder RM, et al. "ITE - Multimedia Information Systems as an Operational interface within the Advanced Manufacturing Environment," Final Report. SERC Grant : GR/H 43038 University of Southampton, 1995.
(4) Microcosm. http://www.multicosm.com/
(5) Crowder RM, Hall W, Heath I, Bernard R, Gaskell D. A Hypermedia Maintenance Information System. IEE Computing and Control Engineering Journal, Vol. 7 No. 3 June 1996 pp121-128.
(6) Crowder RM, McManus P. "Information Live." Manufacturing Engineering, vol 74, pp227- 229. 1995
(7) Davis HC, Hall W, Heath I, Hill G, Wilkins R. Towards an Integrated Environment with Open Hypermedia Systems. Proceedings of the ACM Conference on Hypertext, EHCT'92, Milian, Italy, December 1992, pp 181-190. ACM Press 1992.
(8) Hall W, Davis HC, Hutchings, GA. Rethining Hypermedia: the Microcosm Approach. Kluwer Academic Press, 1996.
(9) Molich R, Nielsen J. Improving a Human Computer Dialogue. Communications of the ACM. Vol. 33 No 3. March 1990 p349-347.
(10) Stobart RK. Design for diagnostic and Condition Monitoring. Conference proceedings, Maintenance, The Business Challenge. I Mech E. Birmingham UK. 16-17 June 1992 pp 109-113.
(11) Crowder RM, Hall W, Heath I, Wills GB. "Requirements Specification: FIRM: Factory Information Resource Management," EPSRC Grant: GR/L/10482. University of Southampton, 1996
