An array processing approach to pilot decontamination for massive MIMO

We address the problem of pilot decontamination for massive MIMO systems for finite dimensional channels, wherein the channel vectors between the users and the base station are composed of a finite number of discrete paths. The pilot decontamination problem is addressed using high-resolution parametric spectral estimation methods, such as root multiple-signal classification (root-MUSIC), for resolving paths based on their angle-of-arrivals. The paths corresponding to the desired user are identified based on their amplitudes and then the paths are used to form the channel estimate. Our simulations show an improved performance in terms of bit-error rate compared to the existing approaches in certain critical scenarios.