Geintra

The main goal of this PhD thesis is to make contributions to the calibration of a beacon-based local positioning systems (LPS). Nowadays, beacon based localization systems are one of the most common solutions used in indoor localization systems. Inside a coverage area several signal emiters (beacons) are placed in fixed positions and by measuring the distance from the beacons to the mobile node it is possible to calculate its position.
In order to achieve the localization of a mobile node, most of the localization algorithms need to know the beacons’ position regarding the reference system that it is being used. Obtaining these positions is often a long and a tedious process, so in most of the papers in the scientific literature the beacons’ position are assumed to be known. But when the real implementation of a local positioning system is carried out, the errors in the determination of the beacons’ position affects to the error in the mobile node localization in a similar way that the errors made in the measurement of the distances to the beacons themselves.
The PhD thesis can be divided into three major parts. In the first one, a new positioning algorithm for hyperbolic trilateration is proposed. After a detailed analysis of the different localization techniques, both spherical and hyperbolic, a new localization algorithm is proposed that is based on the geometric properties of the Cayley-Menger operator. This algorithm gets the mobile node position with a similar accuracy as the classic methods but with a lower computational load. Also a simplified version of the algorithm when all beacons are at the same height is presented. This version further reduces the computational load of the system and avoids the need of a iterative process.
In the second part, several calibration techniques for local positioning systems are proposed. It is studied the case for a single LPS and the case where a global LPS is composed of various independent LPS. The calibration algorithm calculates the beacons’ position taking several measurements from different points of the coverage area. The main contribution of this algorithm is that it only needs to know the actual localization of three measuring points while the rest of the points can be at unknown positions. It is necessary to know the position of three points to define the reference system and to avoid ambiguities in the solution due to rotations and specularities. This method only allows calibrating a single independent LPS system, so a global calibration system for several independent LPS is also proposed. This method uses a mobile robot to calibrate all the LPS systems regarding the same reference origin by merging the odometry data with the LPS systems information through an H-¥ filter.
Finally, in the last part, it is defined a prototype of an ultrasonic local positioning system. This prototype has been used to verify the results and performance of the different algorithms proposed in this thesis.
In order to get the “European Doctor” mention, in the appendix D appears an extended abstract of this thesis in English, including the conclusion section.