An ubiquitous positioning system based on IEEE 802.15.4 radio signals

This work presents a theoretical and experimental study of methods and strategies to implement an ubiquitous positioning system based on IEEE 802.15.4 radio signals. Ranging phase is first analyzed. An empirical method, for a combined use of both Received Signal Strength Indicator (RSSI) and Time Of Flight (TOF) measures, is proposed. Distance measurements performed in different scenarios highlight that RSSI measures provide lower errors on short distance while TOF measures are more suitable for longer distances. The RSSI threshold in use can be different for every scenario but our measurements show that it can be well identified by observing the relevant RSSI calibration curve. Once this threshold is identified, a complementary use of RSSI and TOF measures is then possible. Subsequently the positioning phase is analyzed. The main range-based and centralized localization algorithms are at first compared using simulated localizations, both in their cooperative and non cooperative version, in order to have a complete overview of their performances. However, the considered driving case study (the position monitoring of vehicles/assets in a typical airport area) puts in evidence the requirement for a non cooperative localization strategy. In this condition simulations show that the optimization approach outperforms the other localization algorithms. The deterministic optimization has been also preferred to the stochastic one: position accuracy being equal, the stochastic optimization methods also need the preliminary set up of different parameters through several simulations and trials. Real-world positioning tests, considering indoor and outdoor scenarios of different sizes (from ten to some hundreds of meters) confirm the validity of this choice, the effectiveness of the proposed ranging method and also the importance of exploiting anchors' redundancy. However, position accuracy is still poor for the most noisy scenarios. Future works should focus on the development of an efficient weighting algorithm to further decrease position error.