| PhD Viva

Name of the Speaker: Mr. Gokularam M (EE17D400)
Guide: Dr. Giridhar K
Co-Guide: Dr. Sheetal Kalyani
Online meeting link: https://meet.google.com/nix-tpqd-tus
Date/Time: 11th June 2025 (Wednesday), 6:30 PM
Title: Performance Enhancement in Private Statistical Learning and Distributed Radar Systems.

Abstract :

In this talk, we first consider techniques to improve accuracy in the realm of differentially private estimation and statistical inference. In the era of big data, the paramount concern is no longer the scarcity of information but the protection of our personal details. The framework of differential privacy (DP) renders analysis of the congregated data while ensuring that any individual’s contribution to the dataset remains indistinguishable. We will discuss an additive noise mechanism for DP that adds noise sampled from a new noise density called flipped Huber distribution. It is a hybrid density that resembles Laplace in the centre and Gaussian in the tail. With a sharper centre and light, sub-Gaussian tail, this density has the best characteristics of both distributions and renders a better trade-off between privacy and accuracy than other existing mechanisms. Also, conventionally, independent and identically distributed (i.i.d.) noise samples are added to each coordinate of the query response. We formalize the addition of independent but not necessarily identically distributed (i.n.i.d.) noise across the coordinates to impart differential privacy. 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.

We will then discuss a technique to improve target detection in distributed radar systems with several radiating antennas when there are a large number of targets. 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 talk, a scheme to detect multiple targets in such distributed radar systems will be discussed. 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 that may not be easily detectable. Numerical simulations corroborate the efficacy and usefulness of the proposed method in detecting weak target reflections.