Design and Implementation of a 3D Indoor Localization System Enabling Augmented Reality Applications

This thesis presents design and implement a UWB 3-D real time locating system for real time augmented reality applications to be applied in TV studios and begins presenting the state of art of RTLS taking special care in UWB. The UWB-based Localization System is formed by the Anchors and the Tags. They will be exchanging ranging messages between them to let the tags estimate their position based on a coordinate system. In this thesis, the number of test points is twenty-five and Localization system will be based on indoor area in which 8 anchors are disposed to range with up to three tags that are programmed to work. A measurement campaign has been carry out in RAI by using UWB devices from synchronicIT. This measurement were done with different scenario without obstacle and with human body. For without obstacle the measurements are base on two different tripod 93cm and 153cm and for human body we have three measurements which consist of right shoulder 156cm, right pocket 77cm and back belt 93cm. By calculating ranging error the best performance for scenario of without obstacle and human body will be observed. The localization was effectuated in two phases: ranging and positioning. For the ranging phase was used the time of arrival (TOA) technique and For the positioning phase was used the extended Kalman filter (EKF) since it is robust and less complex than many others algorithms. Generating of EKF filter and state models were done by MATLAB. In this thesis, EKF is simulated to show its features and how its parameters change the tracking performance by using two different state models. In particular, EKF algorithm considers the uncertainty related with the position of the tags that are sending its ranging measurements. This is done by increasing the standard deviation associated to the measurement with a value that depends on the error covariance matrix from the EKF. Several tests were done in order to evaluate and compare the localization performance of different cases without obstacle and human body. It was found that the localization depends on the ranging measurements and with more ranging measurements the system becomes more precise. Any variation of the ranging measurements could affect the localization results. In order to evaluate the effect that the human body might cause to the UWB signal, different tests were performed in RAI TV studio by using EKF algorithm for 1 tag, 2 tags and 3 tags.