Abstract: Integrated sensing and communication (ISAC) is a defining capability of 6G, with 3GPP increasingly treating sensing as a native network function rather than an add-on to communication. Realizing this vision requires estimating target parameters accurately from communication-native OFDM measurements under the strong static and dynamic clutter of real propagation environments.

This seminar presents a body of work on bistatic OFDM-ISAC sensing that addresses target parameter estimation under realistic clutter through complementary learned and model-based approaches. We first develop a coarse-to-fine, learning-based estimator that operates directly on a quantized delay–Doppler–azimuth–elevation (DDAE) representation derived from OFDM pilot measurements, providing sub-grid refinement of delay, Doppler, and angle well beyond grid-limited peak selection, with accuracy that transfers across distinct ray-traced scenes. We then turn to the near-field regime relevant to large-aperture FR3 deployments, where a model-based spatial-focusing and joint spatial–Doppler MUSIC estimator recovers the full 3-D velocity vector in cluttered indoor scenes, together with an analysis of how structured background reflections leak into the signal subspace and degrade estimation.

Together these results outline a path toward deployable, communication-compatible, clutter-robust distributed ISAC sensing, and motivate open directions in multi-target operation and end-to-end multistatic fusion.

Event Details
Title: Sensing for Distributed OFDM-ISAC Networks: Learned and Model-Based Target Parameter Estimation under Clutter
Date: June 23, 2026 at 03:00 PM
Venue: ESB244
Speaker: Mr. Shiv Shankar (EE20D032)
Guide: Dr. Ganti Radha Krishnan
Type: PHD seminar

Updated: