Tuning-Free Online Robust Principal Component Analysis through Implicit Regularization

The performance of the standard Online Robust Principal Component Analysis (OR-PCA) technique depends on the optimum tuning of the explicit regularizers and this tuning is dataset sensitive. We aim to remove the dependency on these tuning parameters by using implicit regularization. We propose to use the implicit regularization effect of various modified gradient descents to make OR-PCA tuning free. Our method incorporates three different versions of modified gradient descent that separately but naturally encourage sparsity and low-rank structures in the data. The proposed method performs comparable or better than the tuned OR-PCA for both simulated and real-world datasets. Tuning-free ORPCA makes it more scalable for large datasets since we do not require dataset-dependent parameter tuning.

Source Enumeration using the Distribution of Angles: A Robust and Parameter-Free Approach

Source enumeration, the task of estimating the number of sources from the signal received by the array of antennas, is a critical problem in array signal processing. Numerous methods have been proposed to estimate the number of sources under white or colored Gaussian noise. However, their performance degrades significantly in the presence of a limited number of observations and/or a large number of sources. In this work, we propose a method based on the distribution of angles that performs well in (a) independent Gaussian, (b) spatially colored Gaussian, and (c) heavy-tailed noise, even when the number of sources is large. We support the supremacy of our algorithm over state-of-the-art methods with extensive simulation results.

Low-Complexity Linear Decoupling of Users for Uplink Massive MU-MIMO Detection

Multi-user massive MIMO is a promising candidate for future wireless communication systems. It enables users with different requirements to be connected to the same base station (BS) on the same set of resources. In uplink massive MU-MIMO, while users with different requirements are served, decoupled signal detection helps in using a user-specific detection scheme for every user. In this paper, we propose a low-complexity linear decoupling scheme called Sequential Decoupler (SD), which aids in the parallel detection of each user's data streams. The proposed algorithm shows significant complexity reduction, particularly when the number of users in the system increases. In the numerical simulations, it has been observed that the complexity of the proposed scheme is only 0.15% of the conventional Singular Value Decomposition (SVD) based decoupling and 47% to the pseudo-inverse based decoupling schemes when 80 users with two antennas each are served by the BS.

Using DCFT for Multi-Target Detection in Distributed Radar Systems with Several Transmitters

In distributed radar systems, when several transmitters radiate simultaneously, the reflected signals need to be distinguished at the receivers to detect various targets. If the transmit signals are in different frequency bands, they require a large overall bandwidth. Instead, a set of pseudo-orthogonal waveforms derived from the Zadoff-Chu (ZC) sequences could be accommodated in the same band, enabling the efficient use of available bandwidth for better range resolution. In such a design, special care must be given to the 'near-far' problem, where a reflection could possibly become difficult to detect due to the presence of stronger reflections. In this work, a scheme to detect multiple targets in such distributed radar systems is proposed. It performs successive cancellations (SC) starting from the strong, detectable reflections in the domain of the Discrete Chirp-Fourier Transform (DCFT) after compensating for Doppler shifts, enabling the subsequent detections of weaker targets which are not trivially detectable. Numerical simulations corroborate the efficacy and usefulness of the proposed method in detecting weak target reflections.

Differential Privacy with Higher Utility through Non-identical Additive Noise

Differential privacy is typically ensured by perturbation with additive noise that is sampled from a known distribution. Conventionally, independent and identically distributed (i.i.d.) noise samples are added to each coordinate. In this work, propose to add noise which is independent, but not identically distributed (i.n.i.d.) across the coordinates. In particular, we study the i.n.i.d. Gaussian and Laplace mechanisms and obtain the conditions under which these mechanisms guarantee privacy. The optimal choice of parameters that ensure these conditions are derived theoretically. Theoretical analyses and numerical simulations show that the i.n.i.d. mechanisms achieve higher utility for the given privacy requirements compared to their i.i.d. counterparts.