Portfolio >
Reconfigurable Factory Testbed Monitoring + Troubleshooting
I worked with a team developing a monitoring + troubleshooting interface for The Reconfigurable Factory Testbed (RFT) at the University of Michigan Engineering Research Center. The RFT investigates flexible manufacturing technologies with remote human management interfaces. The RFT is supported by a grant from the National Science Foundation.
Goals: Our first goal was to develop initial interface requirements, based on user and system research. Following establishment of basic requirements, our goal was to develop a prototype interface + visualizations for monitoring + troubleshooting the RFT.
Challenges: Challenges included establishing a framework for representing changing manufacturing processes, and creating visualizations of data from changing sensor configurations. These needed to show multiple levels of granuarlity + context to support the depth of understanding required by a range of users.
Role: I was one of a team of five UM-SI School of Information graduate students cooperatively designing the RFT interface + visualizations. We frequently worked with the Primary Investigator and other RFT staff. I moved away from Michigan just after completion of the prototype interface.
Outcome: An updated version of the prototype interface is now in active use. The user-based design cycle we established is now a model for ongoing refinement of the RFT interface.
This project followed an iterative design process for developing visualizations and a user interface for the RFT. The iterative stages started with user research, and continued with system+user modeling, design, and testing.
User Research: The majority of our effort was spent on user research. This included observation of RFT staff, auto manufacturing assembly lines, comeptetive analysis of existing systems, and review of literature on troubleshooting complex systems & remote monitoring & control of manufacturing systems.
Modeling: The modeling stage of our design process focused on creating and refining use cases, including detailed task chunking and data exploration processes. The modeling stage often required us to return to the user research stage to gather further information.
Design: The prototype is the direct result of work at the design stage. We focused on creating the design based on our use cases, but put a great deal of effort into producing a prototype that would be flexible enough to work in a much broader range of circumstatnces. The result was a detailed series of screen comps detailing the operation of the interface.
User Testing: User testing was to involve presenting paper prototypes and onscreen animations to RFT research staff and auto assembly line workers. I moved away from Michigan just after user testing began.
Apply Ecological Interface Design:
Ecological Interface Design (EID) provides a framework for integrating conceptually low-level system data with high-level representations of processes, status, and goals. This framework establishes methods for creating flexible visualizations and interfaces for complex systems. A direct result of our application of EID was a smooth path from high level system status information to detailed data analysis.
Use Scenario Based Design:
Scenario based design helps to reduce complex problem areas into managable sets of requirements. We developed scenarios incorporating the extensive user data we collected. These scenarios allowed us to work more quickly, with a tighter focus. However, we remained careful to avoid producing a design narrowly focused on specific scenarios. A direct result of our scenario scenarios was the distinction between the overview interface for operators and troubleshooting interface for electricians + engineers.
Keep it Simple:
Our user research established that the the most frequent users of the interface would be operators with little or no computer experience. Operators needed a straightforward interface that would address their immediate needs and mask much of the system's complexity.
Make Room for Complexity:
Complex systems often behave unexpectedly. An interface involving complex systems should include facilities for working with situations which exceed the capabilities of high-level overviews.
