Engineering & Spatial Telemetry

Decentralized BLE telemetry pipelines, sensor mesh architectures, signal processing algorithms, and open-source scientific tools developed for built environments.

Open Science Artifact Open Source

Passive BLE Presence & Benchmarking Toolkit

Creator & Maintainer

An open-source telemetry processing library for ingesting raw Bluetooth advertisement frames, filtering randomized peripheral MAC addresses, and benchmarking spatial presence heuristics against ground-truth logs.

  • Modular Python telemetry pipeline with zero proprietary cloud dependencies
  • Pre-built Kalman filters and sliding-window median filters to eliminate multipath fading noise
  • Edge-compatible architecture running locally on Raspberry Pi and mini PCs
  • Reproducible algorithmic benchmarks and exportable JSON/CSV logs
Python BLE Telemetry Kalman Filters Edge Computing Open Science
Built Environment Telemetry AVI '26 Demonstration

Campus Spatial Analytics & Mesh Architecture

Co-Researcher & Telemetry Engineer

Architected and deployed a distributed non-intrusive BLE sensor mesh across university campus spaces to monitor occupant density, study room capacity, and mobility trends without capturing personal identifiable data.

  • Distributed low-cost receiver mesh collecting anonymous signal RSSI telemetry
  • Sliding-window RSSI filtering to de-alias randomized MAC addresses
  • Real-time floorplan heatmap rendering and density trend telemetry
  • Privacy-by-design protocol with zero cameras or surveillance hardware
IoT Mesh BLE Telemetry Privacy by Design Data Visualization Edge Telemetry
AVI '26 Paper
Sapienza CS Dept • Transit UX CHItaly '25 Published

Automated Mobility Telemetry & Detection Protocol

Lead UX Researcher & BLE Telemetry Architect

Engineered a passive smartphone BLE beaconing pipeline to detect passenger presence, boarding events, and compartment transitions inside autonomous transit vehicles without QR codes, physical turnstiles, or optical cameras.

  • Built an empirical evaluation testbed inside simulated transit compartments
  • Over 94% passenger boarding detection accuracy across varied pocket/bag positions
  • Ingested multi-receiver signal packets with RF multipath fading suppression
BLE Beacons Python RF Telemetry Field Studies Automated Transit
ACM DOI
Physical Modeling & ML Doctoral Formulation

Sensor Mesh Physics & Spatio-Temporal Graph Formulation

Algorithmic Architect

Formulated the continuous physical dispersion model and Spatio-Temporal Graph Attention Network (ST-GAT) inverting multi-zone environmental telemetry into calibrated occupancy estimates.

  • Governed by mass conservation differential equations modeling inter-zone airflow
  • Dynamic directed graph with learnable multi-head temporal attention weights
  • Evaluated against classical SVM and LSTM baselines with 98.4% accuracy
Graph Neural Networks Spatio-Temporal GAT Differential Equations Scientific ML