arXiv subjects

Thomas Wiese

Publications and source records attributed to Thomas Wiese.

3 recordsLinked to original sources

arXiv2018

A Bilinear Equalizer for Massive MIMO Systems

We present a novel approach for low-complexity equalizer design well-suited for cellular massive MIMO systems. Our design allows to exploit the channel structure in terms of covariance matrices to improve the performance in the face of pilot-contamination, while basically keeping the complexity of a matched filter. This is achieved by restricting the equalizer to functions which are bilinear in the received data signals and the observations from a training phase. The proposed design generalizes several previous approaches to equalizer design for massive MIMO. We show by asymptotic analysis that with the proposed design the achievable rate grows without bound for growing numbers of antennas even in the presence of pilot-contamination. We demonstrate with numerical results that the proposed design is competitive with more complex approaches in a practical cellular setup.

cs.IT↗

arXiv2018

Learning the MMSE Channel Estimator

We present a method for estimating conditionally Gaussian random vectors with random covariance matrices, which uses techniques from the field of machine learning. Such models are typical in communication systems, where the covariance matrix of the channel vector depends on random parameters, e.g., angles of propagation paths. If the covariance matrices exhibit certain Toeplitz and shift-invariance structures, the complexity of the MMSE channel estimator can be reduced to O(M log M) floating point operations, where M is the channel dimension. While in the absence of structure the complexity is much higher, we obtain a similarly efficient (but suboptimal) estimator by using the MMSE estimator of the structured model as a blueprint for the architecture of a neural network. This network learns the MMSE estimator for the unstructured model, but only within the given class of estimators that contains the MMSE estimator for the structured model. Numerical simulations with typical spatial channel models demonstrate the generalization properties of the chosen class of estimators to realistic channel models.

cs.IT↗

arXiv2013

Irregular Sampling of the Radon Transform of Bandlimited Functions

We provide conditions for exact reconstruction of a bandlimited function from irregular polar samples of its Radon transform. First, we prove that the Radon transform is a continuous L2-operator for certain classes of bandlimited signals. We then show that the Beurling-Malliavin condition for the radial sampling density ensures existence and uniqueness of a solution. Moreover, Jaffard's density condition is sufficient for stable reconstruction.

math.NA↗