Path Clustering for Radio Propagation Prediction With Moving Object in Indoor Environment

To simulate radio propagation in a dynamic indoor environment such as a factory operating with moving people and automated guided vehicles (AGV), we devised a technique to determine shielding losses by moving objects based on ray tracing. For a moving object environment, conventional ray-tracing methods require simulation of each snapshot of a moving object. This conventional technique requires too much time for real-time predictions. The proposed method first calculates the static environment using the ray-tracing method without the moving object, and then the shielding loss caused by a moving object is included through post-processing. To enhance efficiency, 3D path clustering was employed, which significantly reduced the total number of paths and computation time. For comparison, we measured the radio-propagation environment of an existing Japanese factory. The root-mean-square errors (RMSEs) and correlation coefficients of the proposed and conventional methods were compared based on the experimental results, wherein the indicators of the proposed method were comparable to those of the conventional method even though the number of paths was reduced. Furthermore, the proposed method reduces the computation time to approximately 5% of that of the conventional method.