LoS Sensing-based Channel Estimation in UAV-Assisted OFDM Systems

TL;DR

This work tackles CE accuracy in UAV-assisted OFDM by exploiting the high probability of LoS paths with a LoS sensing-based CE approach (LoS-EnCE). It introduces a kurtosis-based LoS detector to infer LoS/NLoS scenarios, designs LoS-aid and CFAR-inspired thresholds for resolvable taps, and denoises LS CE accordingly to produce an enhanced CIR. The proposed method yields significant NMSE improvements over conventional LS CE and baseline thresholding, and demonstrates robustness to variations in the LoS probability factor $r$ and the number of resolvable paths $P$. The results highlight a practical, sensing-guided CE framework with potential applicability to 4G/5G UAV-enabled OFDM systems and ISAC-inspired CE paradigms.

Abstract

In unmanned aerial vehicle (UAV)-assisted orthogonal frequency division multiplexing (OFDM) systems, the potential advantage of the line-of-sight (LoS) path, characterized by its high probability of existence, has not been fully harnessed, thereby impeding the improvement of channel estimation (CE) accuracy. Inspired by the ideas of integrated sensing and communication (ISAC), this letter develops a LoS sensing method aimed at detecting the presence of LoS path. Leveraging the prior information obtained from LoS path detection, the detection thresholds for resolvable paths are proposed for LoS and Non-LoS (NLoS) scenarios, respectively. By employing these specifically designed detection thresholds, denoising processing is applied to classical least square (LS) CE, thereby improving the CE accuracy. Simulation results validate the effectiveness of the proposed method in enhancing CE accuracy and demonstrate its robustness against parameter variations.

Figures (4)

• Figure 1: Architecture of the proposed joint processing system.
• Figure 2: NMSE vs. SNR.
• Figure 3: NMSE vs. SNR, where $r=0.4$, $r=0.7$ and $r=1.0$ are considered.
• Figure 4: NMSE vs. SNR, where $P=5$, $P=9$ and $P=13$ are considered.