Indoor Positioning based on Ultra-Wide Band (UWB) Low Cost Technologies: Accuracies and Performances
Paolo Dabove, Vincenzo Di Pietra, Marco Piras, Ansar Abdul Jabbar and Syed Ali Kazim, Politecnico di Torino, Italy
Positioning in indoor environment is one of the most important and interesting topic in navigation system that still present numerous open problems to investigate. Many researchers are attempting to investigate on technologies able to reach valid location in an indoor space, where the well-known GNSS positioning systems are not able to operate. Among the many solutions proposed for location-based services, several wireless communications technologies have the potential to be employed for indoor positioning. In this paper, an Ultra-wideband (UWB) indoor positioning system is presented which exploits two-way time of flight (TWTF) to compute range measurements. These measurements are used in multi-lateration method to compute the position of a trans-receiver (TAG). This kind of system has the advantage to provide high accurate positioning (around 10 cm from the state of art) as well as having low power consumption, high level of multipath resolution, ability to work with low SNRs, high data rate and many more. In particular, the present research is focused on testing and evaluate the performances of commercial solution called Pozyx® with two algorithms i.e. UWB_ONLY (the position is obtained from the ranges) and TRACKING (position is obtained from the integration of range measurements and sensors information). Many sensors like inertial chipsets, magnetometers and pressure sensor are also included with Pozyx® UWB positioning system.
In this paper, quality of the system in term of precision and accuracy has been analysed in different environments. Firstly, the calibration results have been shown, comparing the internal IMU sensor of UWB system with a mass-market one, using the angles measured thanks to the total station as reference.
Secondly, we also have statically analysed the positioning and the ranging capabilities of the system in a favourable environment, as an indoor office room. The average 3D accuracy obtained from the test is 100 ± 25mm. Also the ranging measurements has been analysed as the raw data from which the Pozyx® inner algorithm starts to compute the positions. It has been demonstrated that the range accuracy is about 320 ± 30mm. After these first tests, the system has also been tested in a harsh environment in a narrow corridor where a horizontal accuracy of about 87.4 mm is obtained with a maximum ranging error of ±225 mm. The system and algorithms tested in this report gave almost similar performances in both environments.
In the next steps, other configurations will be tested, considering also more reference anchors in order to increase the accuracy values and the reliability of the solution. From the results of these test, it is possible to state that this kind of system could reach easily a decimetre level of accuracy, a level of accuracy interesting for a large panorama of application. Unfortunately, these systems pretends a network installation in order to provide the position and this could represents a limitation in term of cost, time and infrastructure requirements.