CNN‐based received signal strength prediction: A frequency conversion scheme

To support the design and deployment of emerging high-frequency wireless systems in various applications such as intelligent transportation, advanced received signal strength (RSS) simulation tools for radio coverage prediction are needed. In railway vector parabolic equation (VPE) method is a commonly used approach wave propagation modelling large guiding structures tunnels. However, running VPE simulations at high frequencies an environment computationally expensive, discretization parameters closely related to wavelength. this letter, authors propose convolutional neural network-based model that can efficiently provide RSS by leveraging data obtained from lower frequency anchor. This proposed significantly improve efficiency predictions has been validated against conventional tunnel cases.