Indoor positioning systems have been used as a supplement to provide positioning in settings where GPS does not function. However, the accuracy of calculated results varies among techniques and algorithms used; system performance also differs across testing environments. As a result, users’ responses to and opinions of these positioning results could be different. Furthermore, user trust, most closely associated with their confidence in the system, will also vary. A relatively little studied topic is the effect of positioning variance on a user’s opinion or trust of such systems (GPS as well, for that matter). Therefore, understanding how user interaction with such systems (through trust) changes is important for achieving more usable positioning system design. An experiment was designed to examine if the sequence of location accuracy will affect users’ trust in an individual episode positioning result as well as the system overall. The simulated positioning system running on an iPad used for this experiment provides 10 priming positioning results at a specific category of accuracy. The accuracy is controlled and is presented as either 1. ACCURATE (within 5 meters of actual location), 2. INACCURATE (greater 15 meters), 0r 3. WRONG BUILDING (outside current building’s footprint). After one set of these priming locations a series of 55 post-priming locations across the same categories in addition to 10 CONTINUOUS locations (with between 6 and 15 meters of error) were presented. At each experimental site participants located themselves using the simulated system and rated their trust for that location. Variables obtained from the experiment include: 1. Two types of trust at each location (positioning trust and system trust); 2. Spatial abilities, sense of direction, and ancillary survey data (user characteristics). Results show that users’ trust varies among different accuracy categories and changes over time according to the system performance in association with their own characteristics. Specifically, the accuracy of the priming locations has an impact on users’ trust of later results. Besides, users’ trust in individual positioning results is quite variable and the variability is closely related to accuracy, while user trust of the overall system is less variable.
展开▼
