This study aims to address the problem that Bluetooth Channel Sounding phase observations are prone to distortion in complex indoor multipath environments; this paper proposes a dual-domain signal error compensation method for robust indoor ranging and positioning.
The proposed method integrates frequency-domain phase purification and time-domain first-arrival path extraction. Linear phase bias is first calibrated from repeated observations and nonlinear residual phase errors are modeled to restore purified channel phase vectors. A Dynamic Delay Search Engine (DDSE) is then constructed to constrain Orthogonal Matching Pursuit (OMP) in the delay domain, thereby suppressing coherent multipath pseudopeaks and improving first-arrival path extraction. Multi-base-station ranging results are finally fused through nonlinear least-squares positioning.
Experimental results in a typical indoor scenario show that the proposed method achieves a ranging root mean square error (RMSE) of 0.38 m and a positioning RMSE of 0.60 m. The method outperforms phase-based ranging, inverse fast Fourier transform, MUSIC and iterative Bayesian baselines, demonstrating improved robustness against multipath interference and phase distortion.
This study develops a frequency-time dual-domain compensation framework for Bluetooth CS-based indoor positioning. By combining hierarchical frequency-domain phase compensation with DDSE-constrained sparse delay reconstruction, it provides a practical approach for sub-meter indoor localization under non-line-of-sight multipath conditions.
