The performance of an Indoor Positioning System is highly related to the placement of the transmitting nodes that are used as references for the positioning estimations. Within this graduation project, we propose a methodology that can be used to optimize such a deployment and thus, increase the performance of an Indoor Positioning System that a) is based on Received Signal Strength Fingerprinting and b) is orientated towards providing location or zone estimations instead of exact positioning. The optimization process involves 4 fundamental components. Firstly, the modeling of the obstructions in the indoor environment and also the zone modeling. Then, the definition of the performance metric that can be used to evaluate each different deployment scenario, in which case, our proposed metric considers the separation area and distances between the zones in the RSS vector space. The third component is the radio propagation model, required for simulating the transmitted signals from each node, where a model based on the ray tracing technique is selected. Finally, the last component is the selection of the optimization function that will control and drive the whole optimization process by choosing which deployment schemes to evaluate. For that, the utilization of a Genetic Algorithm has been selected. The evaluation of our methodology showed that the most problematic regions in terms of localization accuracy are, as expected, those where different zones become adjacent. Yet, comparisons between regular node deployments and our optimized solutions indicated that, regardless the number of nodes, our optimization introduced in each case an overall localization improvement that was especially concentrated at the most problematic regions.

This report constitutes the final product of my internship, where research was conducted on the differences of the Bluetooth Low Energy signal reception between 2 different phones (p1, p2) and eventually, the possibility of developing a translation function that could be used to predict the signal strength reception of p1, by considering the signal strength reception of p2. Such model would be particularly useful to applications related to Indoor Localization/Positioning, as these are often based on the BLE signal strength. For the development of this model, the influence of several parameters was assessed, such as: a) the distance between a phone and a beacon, b) their orientations and c) the number of concurrently broadcasting beacons, and all were found to be significant. Furthermore, it was discovered that as long as there is no movement in the system, the BLE signal reception at a specific channel has low variations and so, even a few samples can be representative for each channel. The evaluation of the translation functions was quite promising. Ultimately, by taking advantage of a specific Android's behaviour during the training phase, it became possible to identify the channels of incoming BLE signals. This information was then used to significantly enhance the performance of the translations under specific circumstances (i.e. the channels can be identified during the operational phase too).

We are now gradually entering the era of big data - maybe a bit too much of a buzzword, but it is not lied. Technology is evolving fast, enabling faster and more efficient data acquisition, storage, retrieval and processing. Point cloud datasets are such a type which relies on large files and lots of processing power. The rather fast evolutions in technology enable the shared idea between Delft University of Technology and Fugro of an ‘Open Point Cloud Map’. This Open Point Cloud Map aims at making point cloud datasets easily available to the public, even letting them performsimple analysis. Both Fugro and TU Delft want to take lead in development of such an environment; three student teams from TU Delft thus form a partnership with Fugro to kick-off three in-depth researches which would result in one step closer to the vision of OPCM. The ChronoCity team will focus on the time-component (Figure 1.1) in the acquainted point clouds, while the other two teams focus on location-dependency and different scales and granularities of the datasets.

The Chronocity-team strives to create an online interactive tool which gives the user the ability to view, explore and analyze massive point cloud datasets on-the-fly. Since the limited timespan in which the project should take place this would not yield a fully optimized application, but at least the general principles are defined and evaluated on for a more defined future in the development of the OPCM. A large portion of the efforts will go into making the data and analyses available to the public - in an interactive and user-friendly way; because without this availability, the underlying principles are not brought to the public also. Regarding these underlying principles the most important one is change detection. During the project a suitable algorithm is designed and evaluated for detecting new, removed and changed geometric points.