Definición de emplazamientos potenciales RAN mediante técnicas de inteligencia computacional

Abstract

The project addresses the analyses required by Radio Frequency engineering to define the location of potential sites for the Radio Access Network, using computational intelligence techniques, with the aim of developing a software that allows analysing the most relevant performance indicators (KPI) , define its level of potentiality, generate clusters with such samples to finally obtain the possible locations where the operators will require new base stations. The development of the application was carried out by means of the rapid evaluation of prototypes until reaching an optimal performance, through successive stages: First a homogenization of the sample, second evaluation by means of fuzzy logic of the KPIs, third grouping of the samples using networks neuronal and fuzzy clustering. Once the project objectives had been achieved, a functionality was also implemented to obtain multi-operator sites. Using two real scenarios, the results obtained through the application were evaluated with those obtained by two professionals in RF engineering. At the end of the exercise, both coincidences and differences were observed, which led them to rethink the exercise and finally accept that the analysis of the tool is both deep, accurate and fast. At present, such analyses are very demanding in time and effort, as it is a totally manual analysis process that involves a large amount of information. With the developed tool, immediate results were obtained, the most important variables were analysed in their entirety and the possibility of obtaining sites with multi-operator potential was also obtained.