A Realizable Physics-Corrected LCCR-UW-OFDM Framework with Blind Channel-Order Estimation and Adaptive Reliability-Gated Iterative Refinement
Year: 2026
Type: article
Abstract: Orthogonal frequency division multiplexing with unique word (UW-OFDM) provides a guard interval that is deterministic in nature, offering better spectral efficiency and lower out-of-band emissions than conventional standard Cyclic Prefix structures (CP-OFDM). Nevertheless, in severe time dispersive and frequency selective multi-path channels, deterministic unique word sequence is affected by physical time distortion. The resulting propagation creates a harmful tail leakage following the user equipment (UE), which spills over into the active data window and destroys the orthogonality of the subcarriers leading to irreducible ICI floors. To address this fundamental limitation, we propose a realizable, physics-corrected Leakage Compensation and Cyclic Restoration (LCCR) framework with an Estimated UW Interference Reconstruction (EUIR) module. The suggested framework functions without any prior "genie-based" information, unlike leading-edge deep learning models or matrix-inversion frameworks. This feature employs a threshold for structural model-order selection to make an accurate estimate of the active physical channel memory order from raw time-domain energy profiles. Also, a reliable-gated decision-directed (DD) feedback loop dynamically extracts high-confidence data subcarriers as virtual pilots and prevents error propagation across iterations of tracking. Through heavy Monte Carlo simulations in 8-tap Rayleigh fading channels, it is confirmed that the suggested LCCR compensator has 25% lesser channel estimation error compared to existing systems, having an MSE of 1.8755 in heavy -5” dB” scenario. Due to its balanced computational complexity bounded by O(N log_2N ), the framework can be optimally utilized in next-gen 6G low-latency wireless edge receivers.
Authors: Mohammed Alkubaisi, Ameer Kadhim Hashosh
Language: English
Subfield: Electrical and Electronic Engineering
Field: Engineering
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7235470471
