The prediction of the electric (E) field plays an important role in the monitoring of the radiofrequency electromagnetic field (RF-EMF) exposure induced by cellular networks. Novel approaches, such as artificial intelligence based approaches are being exploited in the exposure mapping. With the help of public accessed datasets, i.e., cartoradio and OpenStreetMap, we can extract features, related to propagation model.
The risk perception of electronic-magnetic field (EMF) exposure is nowadays a hot issue with the fast development of wireless communications. Usually, RF-EMF monitoring is often carried out using “one-time” measurement campaigns. Measurement equipment includes spectrum analyzers, exposimeters, and network-based mobile phone tools. The monitoring can also be done using fixed sensors installed and tested in cities such as Paris. Besides in situ measurements, statistical methods, e.g., ray-based simulators, and Kriging, are used in assessing EMF exposure. In this work, we did assessments and forecasting of EMF exposure for both outdoor and indoor environments by using ANN based on data collected by simulations, sensors measurements, network-based measurements.
The performance of wireless networks is fundamentally limited by aggregate interference, which depends on the spatial distributions of the interferers, channel conditions, and user traffic patterns. Empirical evidence suggests, however, that practical cellular network deployments are likely to exhibit some degree of interactions among the locations of the BSs, which include spatial inhibition, i.e., repulsion, and spatial clustering. Conventional Poisson point process based approaches are not sufficient anymore. For this, accurate system-level performance characterization and evaluation with spatio-temporal correlation are required. In this project, we studied the performance of spatially-correlated cellular networks, including coverage probability, MISR, Meta distribution, and aggregate average rate under a multi-operator sharing scenario.