An adaptive channel prediction algorithm based on fractal Brownian motion model

Abstract

Adaptive channel prediction (ACP) technology is a promising tool for improving the spectral efficiency on time-varying mobile channels while keeping a predictable bit error rate (BER). In this paper, we describe the fractal Brownian motion (FBM) model and get the interior affiliation between the multi-path fading and FBM model by comparing various fractal characteristics of fading signals, especially the existence of self-similarity. The self-similarity of fractal characteristics indicates the existence of a nontrivial predictive structure, where the fractal dimension and variance are important parameters for describing the signal propagation. The wavelet synthesis algorithm based on FBM is proposed to reconstruct multi-path signals, yielding reasonably accurate replication at a cost of allowable calculating error. Simulation results indicate that the FBM model could predict wireless channel more accurately and effectively than those statistical models on the minimum SNR.