Robust Tensor Positioning Based on Channel Parameter Estimation Under Spatially Colored Noise

In communication systems, the ubiquitous time-frequency interference, such as co-channel interference (CCI) and RF front-end interference (FEI), constitutes spatially correlated colored noise. They can not only result in increased bit error rates and reduced system capacity, but also introduce distortions and errors in phase estimation. Consequently, for the channel impulse response (CIR) testing, the separation of interference components has become a prominent research focus. This paper deduces a complex-domain robust variational Bayesian method for channel parameter estimation, along with the estimation of interference locations. Simultaneously, by using variational mean-field theory and automatic rank reduction, it achieves the automatic determination of the number of multipath and low-rank approximations of the CIR. Finally, the channel harmonic parameters are employed in the angle positioning algorithm. Performance comparisons are made with traditional information-theoretic rank estimation methods and the space alternating generalized expectation maximization (SAGE) parameter estimation algorithm, confirming the feasibility of the proposed approach.