Uncertainty Management in Context-Aware Applications: Increasing Usability and User Trust

Abstract

Uncertainty in context-aware computing is mainly a consequence of the complexity of context acquisition mechanisms and context processing. The presence of uncertainty may harm the users' confidence in the application, rendering it useless. This paper describes a three-phase strategy to manage uncertainty by identifying its possible sources, representing uncertain information, and determining how to proceed, once uncertain context is detected. We conducted an experiment to assess the impact and utility of uncertainty management in a location-aware museum guide that incorporated the approach proposed. The impact was measured by comparing the number of erroneous location estimations and the number of interactions performed by users. A survey was used to measure user friendliness, the utility of the application and user trust among 118 museum visitors who participated in the experiment. Results suggest that uncertainty management helps improve the accuracy of context estimates and the user's perception on the reliability of the application.