Uncertainty Management in a Location-Aware Museum Guide

Abstract

Uncertainty management remains a key issue in the design of reliable context-aware applications. Uncertainty is originated by both, the complexity associated to the acquisition of primary contextual information and imprecision in the derivation of secondary context. If uncertainty is not adequately indentified and managed, it could render context-aware applications useless to the user. In order to assess the impact and utility of uncertainty management in a location-aware application, we conducted an experimental evaluation of a location-aware electronic guide for a museum. The experiment had two conditions: the first one directly uses location information estimated from the strength of the RF signal received in the mobile device from nearby WiFi access points, while the second implements a context uncertainty management mechanism being proposed. The impact was measured by comparing the number of location estimation errors and the number of user interactions required to operate the system. A questionnaire was used to measure the users’ perception on ease of use, utility and trust on the application. The system was evaluated by 118 randomly selected visitors to the museum. Results suggest that uncertainty management helps improve the accuracy of the context estimate and the application performance. We also found that participants considered the implementation of context-aware technologies useful to assist them during their visits to the museum, regardless of the version of the application they used. However, ease of use and user trust is affected by the presence of uncertainty in the contextual information used by the application.