Perspectives Using a Reinforcement Learning Approach and Ray-Tracing SW for 5G+Indoor Coverage Optimization

In wireless communications, measurement-based stochastic GSCM model, ray-tracing based deterministic model, and hybrid model are used. The prediction method based on ray-tracing provides propagation models for environments. Also, material properties were modelled and applied to a ray-tracing analysis. This paper presents an extended reinforcement learning approach for the deterministic ray-based propagation method for indoor environments. To cope with the explosive use of wireless communication, we are studying a method of applying RL to the propagation model of a deterministic prediction method to optimize antenna location combinations and coverage extensions in indoor scenarios such as small cell and DAS systems.