| Resumo: | More and more, mobile applications require precise localization solutions in a variety of environments. Although GPS is widely used as localization solution, it may present some accuracy problems in special conditions such as unfavorable weather or spaces with multiple obstructions such as public parks. For these scenarios, alternative solutions to GPS are of extreme relevance and are widely studied recently. This dissertation studies the case of EduPARK application, which is an augmented reality application that is implemented in the Infante D. Pedro park in Aveiro. Due to the poor accuracy of GPS in this park, the implementation of positioning and marker-less augmented reality functionalities presents difficulties. Existing relevant systems are analyzed, and an architecture based on pedestrian dead reckoning is proposed. The corresponding implementation is presented, which consists of a positioning solution using the sensors available in the smartphones, a step detection algorithm, a distance traveled estimator, an orientation estimator and a position estimator. For the validation of this solution, functionalities were implemented in the EduPARK application for testing purposes and usability tests performed. The results obtained show that the proposed solution can be an alternative to provide accurate positioning within the Infante D. Pedro park, thus enabling the implementation of functionalities of geocaching and marker-less augmented reality.
|
|---|