Signal Processing & RF Ecosystem

Experimental projects exploring SDR spectrum monitoring and biometric signal processing.

Software Defined Radio

Illumination Reflection Tracking

Phase 2/4 (56%) · Full Details →

What It Is: Building on open source implementations to analyze reflections of existing broadcast transmissions using coherent multi-channel SDR hardware.

Current Status: Phase 2 (Clean, stable foundation for reliable tracking)

Tech Stack: Python, numpy, scipy, Rust (planned migration)

Project Spectra — SDR Client

Active Development · Full Details →

What It Is: A high-performance SDR client for spectrum monitoring and automated signal discovery. Spectra transforms raw IQ samples into a “Signal Census” using real-time ML classification.

Real-Time Visualization: The Spectra interface provides both high-level signal census data and low-level waterfall analysis for deep-dive signal inspection.

Spectra Signal Census: Automated detection and classification of signals across the monitored band.

High-resolution waterfall display showing temporal signal patterns and frequency hopping.

Current Status: Active development of ML classification engine.

Tech Stack: Python, numpy, scipy, RTL-SDR, HackRF, PyTorch (classification)

Wi-Fi Radar — Through-Wall Detection

Active Development · Full Details →

What It Is: Using coherent SDR arrays for through-wall human detection via Wi-Fi signal phase analysis.

Current Status: Proof-of-concept detection working

Tech Stack: Rust

rtltcp-rust — SDR Network Streaming

Active Development · Full Details →

What It Is: A high-performance Rust server that streams raw IQ samples from multiple SDR devices (RTL-SDR, AirSpy HF+) over the network using the industry-standard rtl_tcp protocol, with a built-in TUI for live configuration.

Key Features:

• Multi-threaded architecture for concurrent streaming from multiple SDRs
• Terminal User Interface (TUI) for real-time frequency, gain, and sample rate adjustments
• TOML-based persistent configuration
• Cross-platform with Raspberry Pi target for headless remote stations

Current Status: Active development of v1 core streaming and hardware support

Tech Stack: Rust, tokio, ratatui (TUI), librtlsdr/libairspyhf (FFI)

Audio Processing

soundarray — Spatial Audio Processing

Active Development · Full Details →

What It Is: An exploration-focused audio processing system using Raspberry Pi and microphone arrays. Focuses on spatial audio (ToA, beamforming) and classification (vehicles, wildlife) using an “analyst” agent approach.

Key Features:

• Time of Arrival (ToA) Estimation: Localize sound sources using microphone array phase differences
• Beamforming: Directionally filter audio to enhance signals from specific angles
• Sound Classification: Classify vehicles and wildlife by acoustic signatures
• Agent Integration: Provides structured insights to analyst agent frameworks

Current Status: Exploring hardware options (ReSpeaker, Matrix arrays) and beamforming algorithms

Tech Stack: Python, numpy, scipy, PyTorch/TensorFlow (classification)

Health & Biometrics

HealthyPi Biometric Signal Processing

Experimental · Full Details →

What It Is: Experimental signal processing using the HealthyPi biometric hardware platform (developed by Protocentral) for ECG, PPG, and respiration analysis with NeuroKit2.

Current Status: Phase 6 (NATS integration + reconnection handling + tests)

Tech Stack: Python, NeuroKit2, numpy, scipy, NATS

Open Source & Contributions

• HealthyPi Ecosystem: github.com/sk2/healthypi
• Project Spectra: github.com/sk2/spectra
• rtltcp-rust: github.com/sk2/rtltcp-rust
• soundarray: github.com/sk2/soundarray
• Illumination Reflection: github.com/sk2/passive-radar
• Wi-Fi Radar: github.com/sk2/wifi-radar