Robust Ultra-Wideband Direction Finding in Dense Cluttered Environments

Ultra-wideband impulse radio (UWB-IR) exhibits good immunity to the multipath in dense cluttered environments because of its high temporal resolution. However, UWB-IR source direction finding is more challenging due to the frequency-selective distortion. Furthermore, the multipath interferences in both far-field and near-field regions affect the direct path resolution. This paper presents a robust solution for direction finding using UWB-IR in the presence of far-field and near-field multipath components overlapping with the direct path signal. An alternating projection (AP) method is proposed to separate the multiple interfering plane waves for better angle-of-arrival (AOA) estimation accuracy based on the signal subspace of UWB discrete Fourier transform (DFT). Detection of the number of plane waves is suggested before AOA estimation for fast computational convergence. Compared to the MUSIC algorithm based on IEEE 802.15.4a channel simulation, the AP has average root mean square error (RMSE) of less than 2.3 degrees while the MUSIC has poorer RMSE of 22.6 degrees in angular resolution when the signal-to-noise ratio (SNR) ranges from 41 dB to 15 dB. The AP is also found to perform well in the presence of strong multipath overlapping with the direct path with AOA average RMSE error of less than 3.4 degrees.