This project implements a real-time passive radar system using the RTL-SDR platform and Python. The radar passively receives signals (e.g., from FM broadcast stations) and computes target motion using range-Doppler analysis. The application is visualized through a PyQt5 GUI and uses multithreading, real-world units, multi-antenna support, target tracking, clustering, and real-time database logging.

• Why: Prevents GUI freezing while SDR samples are processed.
• How:
  • Each active surveillance antenna is assigned a RadarWorker thread.
  • These threads run signal processing routines (correlation, FFT, range-Doppler mapping) asynchronously.
  • Results are emitted via PyQt signals and handled in the main thread for GUI-safe updates.

• Why: Allows comparison between receivers, directional tracking, or spatial diversity.
• How:
  • Up to 4 antennas are supported.
  • Each has a checkbox in the GUI to toggle it on/off.
  • Each active antenna creates its own RtlSdr object and thread.

• Why: Organizes visualization per antenna for clarity and scalability.
• How:
  • A QTabWidget holds tabs labeled "Antenna 1" through "Antenna 4".
  • Each tab contains a Matplotlib canvas showing a range-Doppler map.

• Why: Enables detection of objects based on delay (range) and Doppler shift (velocity).
• How:
  • Cross-correlation is used to find range.
  • Doppler FFT is applied across a sliding window of correlated samples.
  • Units:
    • Range axis in meters
    • Velocity axis in meters per second (m/s)

• Why: Groups peaks in the range-Doppler map into targets, reducing noise and improving tracking.
• How:
  • Peaks are identified where power is within 10 dB of max.
  • DBSCAN is applied to spatially cluster these into grouped detections.
  • Centroids of these clusters are passed to the tracker.

• Why: Maintains continuity of moving objects and assigns persistent identifiers.
• How:
  • SimpleTracker maintains a list of tracks with positions, history, and a unique ID.
  • Tracks are updated frame-by-frame based on proximity to current detections.
  • Tracks that disappear for a defined number of cycles (max_lost) are pruned.

• Why: Translates sample data into real-world interpretable values.
• How:
  • Range in meters using: range_bin * c / (2 * sample_rate)
  • Velocity in m/s using: doppler_freq * λ / 2 (λ = wavelength)

• Why: Enables persistent identification across updates and sessions.
• How:
  • Each new target is assigned an id when initialized.
  • The ID is displayed next to the target in the plot and stored in the database.

• Why: Keeps a persistent history of target activity for analysis or export.
• How:
  • A SQLite database is initialized (radar_tracks.db) with a tracks table.
  • Each tracked object (range, velocity, time, antenna, ID) is logged in real time using INSERT INTO.

• passive_radar_gui.py: Main GUI + processing logic

• Python 3.8+
• RTL-SDR compatible hardware
• Dependencies:
  • PyQt5, numpy, scipy, matplotlib, scikit-learn, pyrtlsdr

pip install PyQt5 numpy scipy matplotlib scikit-learn pyrtlsdr

python passive_radar_gui.py

• Add heading estimation or AoA (angle of arrival)
• Network streaming of data
• Advanced tracking (Kalman filter, JPDA, etc.)
• Web interface for remote viewing