Robotic RF Localization & Experimental Systems

projects

I lead the architecture and development of MobiFR3, a heterogeneous 10-GHz robotic RF testbed for localization experiments and SLAM-based navigation. The current system connects Pi-Radio SDR channel sounding with TurtleBot4/ROS 2, odometry, 2D LiDAR, and RGB sensing. Jackal UGV integration is in progress.

My role spans system architecture, experiment design, RF-robot synchronization, measurement-aligned simulation, localization inference, and evaluation. I mentor student collaborators in RF hardware operation, calibration, and data collection; routine RF acquisition is conducted as a team rather than presented as a single-person hardware effort.

MobiFR3 RF-enabled TurtleBot4 platform

MobiFR3 integrates the mobile robot, RF front end, compute, power, and control layers into one experimental platform.

Pi-Radio SDR hardware bench

The team-operated Pi-Radio SDR stack combines FR3 RF hardware, Vivaldi antennas, and local compute/control equipment.

From system to inference

MobiFR3 provides the measured data and controlled geometry used by MAGNETAR. The algorithm learns a joint posterior over transmitter position and heading from asynchronous AoA/SNR multipath snapshots, with real-to-sim calibration based on measured antenna patterns and RF-chain/noise effects. This connects the physical system directly to probabilistic inference rather than treating the robot as a demonstration platform alone.

Media & Public Demonstrations

Interview recorded at the Brooklyn 6G Summit 2025.

At the 2025 Brooklyn 6G Summit, I demonstrated the FR3 robotic sensing and localization system, combining TurtleBot4 mobility, Pi-Radio hardware, angle-of-arrival measurements, and SLAM-based mapping. The demonstration showed how RF observations can support indoor positioning when visual information is incomplete or unavailable.

Watch the interview 5G Technology World feature NYU Tandon coverage

System capabilities

  • Pi-Radio SDR channel sounding with Vivaldi antennas and motion-controlled 10-GHz measurements.
  • Transmit/receive waveform control, capture, synchronization, channel estimation, SNR estimation, and AoA processing in a team-operated workflow.
  • Scripted TurtleBot4, linear-track, and D48 pan-tilt motion for reproducible measurement grids.
  • Measurement-aligned Sionna RT scenes with TX/RX pose and motion, measured antenna patterns, and stochastic RF-chain/noise effects.
  • Physical validation for posterior localization and future closed-loop sensing/navigation experiments.
Wireless Robotics Systems