Wavelet-Based Channel Modeling for OFDM-Integrated Pinching Antenna Systems

Abstract

This paper studies wavelet-based channel modeling for orthogonal frequency-division multiplexing (OFDM) links with pinching antenna systems (PASS). A Daubechies-4 decomposition of the PASS channel response gives 4.76 percent sparsity in the transform domain, used for compressed-sensing-based channel estimation. The level-three approximation coefficients carry about 92 percent of the signal energy and mainly correspond to waveguide dispersion. The level-two and level-three detail coefficients are linked to pinching-antenna coupling, while the level-one detail coefficients capture free-space fading and noise. A relative change of up to 20 percent in the phase-matching parameters changes the normalized mean-square error (NMSE) by less than 0.25 dB, so the method does not depend on exact phase matching. The wavelet representation reduces cyclic-prefix overhead from 10.2 percent to 6.1 percent, while orthogonal matching pursuit (OMP) based recovery gives about forty times lower complexity than minimum mean-square error (MMSE) estimation. Together with the 15 percent sparse-denoising gain, the net throughput improvement is about 20 percent. The overlap factor of 0.4 satisfies the theoretical bound, and the rate gains are reported conservatively.