Project: PRECARN 07B — Ecological Interface Design for the Intelligent Drinking Water Monitoring System
|Partner:||Precarn Inc./A.U.G. Signals, Sustainable Development Technology Canada|
|Dates:||January 2007 - January 2009|
|Participants:||Jamieson, G., Vary, B., Kan, K.|
|Collaborator:||A.U.G. Signals Ltd., Toronto Water and Wastewater Services – City of Toronto, Queen’s University, Public Health Agency of Canada, University of Calgary|
|Description:||Contamination events in water distribution networks are difficult to detect in a timely manner using current methods, which rely on manual sampling and laboratory testing. To provide earlier contamination warnings, the Intelligent Drinking Water Monitoring System (IDWMS) automatically samples water from distribution networks and analyzes it for anomalies.|
The goal of this project was to design a graphical display for the IDWMS that helps water distribution operators understand and use the system to detect contamination events. The Ecological Interface Design approach was used to model the IDWMS and identify the necessary information content for the display, and to present the information using appropriate graphic forms.
The first phase of the project also included a Control Task Analysis that was based on field observations and interviews with operators at a water utility control centre. The resulting display prototype, which presented IDWMS sensor readings and detection results, was evaluated with operators in a usability study.
The second phase of the project revised the prototype to provide more detailed information about the system’s capability in order to foster appropriate operator reliance. A visualization for the IDWMS’s sensor fusion algorithm (Dempster’s rule of combination) was created to help operators identify situations in which the algorithm could produce misleading results, This graphical representation of Dempster’s rule was evaluated in a signal detection task performed by university students, and was found to have little benefit in supporting appropriate reliance on the algorithm’s results, Results of this researt≥ ch may help determine the applicability of ecological displays to helping operators appropriately rely upon decision aids in their decision-making.