01 · DroneSAR

UAV-Based Synthetic Aperture Radar

DroneSAR is a compact S-band synthetic aperture radar platform designed for flexible, high-resolution measurements at agricultural field scale. I lead the end-to-end development effort, from system integration and flight experiments through signal processing and science-product generation.

  • S-band FMCW radar imaging from a multirotor UAV platform
  • Polarimetric calibration, motion compensation, and SAR image formation
  • Georeferencing and generation of calibrated, analysis-ready radar products
  • Field campaigns connecting radar observations with in-situ soil and vegetation measurements
DroneSAR UAV-based synthetic aperture radar system

DroneSAR platform photo and demo video.

02 · Retrieval

Soil Moisture & Agricultural Remote Sensing

A major goal of DroneSAR is to turn high-resolution backscatter measurements into useful geophysical products. My research develops and evaluates soil-moisture retrieval methods across changing soil roughness, precipitation, and vegetation conditions, with an emphasis on rigorous field validation.

  • Physics-based and empirical soil-moisture retrieval
  • Bayesian estimation and uncertainty-aware retrieval methods
  • Time-series analysis across wetting and dry-down events
  • Extensions toward biomass and other agricultural products
Example soil moisture retrieval map from DroneSAR

Example field-scale soil moisture map from UAV SAR observations.

03 · SoOp / GNSS-R

Signals of Opportunity for Earth Observation

I have contributed to research using existing communication and navigation transmissions as illumination sources for remote sensing. This includes work related to SNOOPI, a CubeSat demonstration of P-band signals-of-opportunity remote sensing, and participation in the CYGNSS research community.

  • Bistatic and passive microwave remote sensing
  • GNSS reflectometry and signals-of-opportunity concepts
  • Mission operations and science applications for small-satellite sensing
  • Soil moisture and land remote sensing across multiple microwave bands
SNOOPI CubeSat mission image

SNOOPI mission concept for P-band signals-of-opportunity remote sensing.

04 · Estimation

Navigation & Multi-Sensor Fusion

My navigation work focuses on robust estimation when individual sensors become intermittent, biased, or locally accurate but globally drifting. I am particularly interested in architectures that combine visual-inertial information with GNSS and other absolute measurements.

  • Visual-inertial odometry (VIO) and GNSS integration
  • Kalman filtering, smoothing, and factor-graph formulations
  • Error-state and stochastic sensor modeling
  • Performance evaluation under degraded sensing conditions

Current emphasis

  • Robust positioning in real-world environments
  • Architectures that blend relative and absolute sensors
  • Evaluation workflows for drift, bias, and degraded sensing
  • Bridging estimation theory with deployable systems

This area connects my academic work with applied industry-facing positioning problems.

05 · Industry

Internship / Corporate Project Experience

In summer 2026, I joined Qualcomm as a Location Technology Intern. The public description here is intentionally high-level, but the work aligns with my broader interests in positioning, navigation, estimation, and sensor fusion.

  • Applied positioning and sensor-fusion workflows
  • Analysis of system performance under challenging sensing conditions
  • Use of visual-inertial, inertial, and absolute measurements in location pipelines
  • Research-to-product thinking for deployable location technology

Because this is industry work, proprietary details are not included on the public site.

Applied research · Summer 2026

Location Technology

The same estimation questions appear across remote sensing and navigation: how to combine imperfect measurements, represent uncertainty, control drift, and turn raw observations into reliable products. My internship complements my academic work by applying these ideas in an industry location-technology setting.

Public description only; proprietary implementation details are intentionally omitted.