Image Structural Similarity-Based Channel Similarity Analysis across Multiple Frequency Bands in the Office Campus Environment.

The coexistence of multiple frequency systems operating in well-separated frequency bands show great promise for achieving both breadth and depth of coverage. It is necessary to investigate frequency-dependent channel characteristics to facilitate multi-band collaborative communications. By treating the channel power spectrums as two-dimensional images and comparing the structural similarity (SSIM) between the images, a channel similarity index named weighted SSIM (WSSIM) is proposed for a evaluation of the similarity degree numerically. A power-dependent weight is introduced to reduce noise interference to similarity judgment. A large amount of simulation experiments have been conducted across multiple frequency bands in the outdoor office campus environment to obtain channel data. The statistical distributions of the WSSIM, which reflect the similarities between channels of different frequencies, are also presented.