Local positioning system with ultrasonic beacons for 3D environments
- Mannay, Khaoula
- Jesús Ureña Ureña Director
- Taoufik Aguili Co-director
- Mohsen Machhout Co-director
- Álvaro Hernández Alonso Co-director
Defence university: Universidad de Alcalá
Fecha de defensa: 24 February 2021
- Carlos Manuel de Marziani Chair
- Juan Jesús García Domínguez Secretary
- Taher Ezzedine Committee member
Type: Thesis
Abstract
This PhD Thesis contributes with the development and improvement of Indoor Location and Positioning Systems (ILPS), which are used to locate, position and track people, as well as mobile and/or connected targets, such as robots or smartphones, not only inside buildings in the lack of GNSS (Global Navigation Satellite Systems) signals, but also in constrained outdoor situations with reduced coverage. Indoor positioning applications and their interest are growing in certain environments, such as commercial centres, airports, hospitals or factories. Several sensory technologies have already been applied to indoor positioning systems, as infrared, Wi-Fi, light, cameras, or radiofrequency, where ultrasounds are a common solution due to its low cost and simplicity. This thesis deals with the development of 3D positioning systems based on ultrasounds. So, its contributions are divided into three blocks. The first one proposes a 3D Ultrasonic Local Positioning System (ULPS), based on a set of three asynchronous ultrasonic beacon units, capable of transmitting coded signals independently, and on a 3D mobile receiver prototype. The proposal is based on the aforementioned beacon unit, which consists of five ultrasonic transmitters oriented towards the same coverage area and has already been proven in 2D positioning by applying hyperbolic multilateration. Those beacon units are manually calibrated and placed in strategic and known positions of three perpendicular walls (generally in the centre of the ceiling and two perpendicular walls). This approach has been characterized and experimentally verified, trying to maximize the coverage zone, at least for typical sizes in most common public room and halls. The second block deals with several fusion methods, to obtain the final estimated position of the mobile receiver existing inside the positioning space, assuming a low accumulative error, after merging the particular results from each beacon unit. Two merging ways have been presented and implemented: the loosely and the tightly coupled fusions. For the loosely coupled method, three algorithms have been applied: the Maximum Likelihood Estimation (MLE) fusion algorithm, the Linear Kalman Filter (LKF) and the Adaptive Kalman Filter (AKF). These algorithms fuse the positions obtained from several ULPSs to get a final more accurate position. With regard to the tightly coupled fusion methods, three algorithms have also been applied, which are based on: the Extended Kalman Filter (EKF) for only one ULPS; three EKFs for the three independent ULPS; and finally only one EKF for all the set of three ULPSs. On the other hand, the third block proposes a preliminary SoC architecture based on a FPGA device for the receiver stage, so it can be deployed on board a mobile target (people, robot, drone, smartphones, etc.). The architecture involves a specific hardware peripheral, connected to the processor, which is in charge of implementing the low-level processing of the ultrasonic signals (particularly a BPSK demodulation and a transmission encoding with Kasami sequences). Finally, all the proposals aforementioned have been verified by simulations and experimental tests, contributing to the design and improvement of the ultrasonic LPSs as well as to the deployment of these systems in several real environments. Simulations and experimental tests have been satisfactory, achieving a positioning accuracy in the range of centimetres in the zone where the coverages from the three ultrasonic beacon units are available, whereas it is in the range of decimetres whether the coverage from one or more beacon units is missing. Particularly, two different experimental environments have been considered: a small volume with many furniture (Lab), and a large and empty volume (Hall); tests have been carried out at points distributed in the environment to consider those cases of interest.