libgnss-recv-gps

Embedded System Friendly GPS Receiver by Orientable AI

A high-performance GNSS library and application provided by Orientable AI.

📖 Documentation

The full documentation, including installation guides, configuration manuals, and architectural overviews, is available at:

👉 Read the Documentation (Or build the MkDocs site locally)

⚡ Quick Start

1. Build

mkdir build && cd build
cmake ..
make -j$(nproc)

2. Run

# Run headless (CLI)
./orientable-libgnss-cli ../cfg/orientable-libgnss.yaml

# Run modern GUI (Qt6 or GTK4)
./orientable-libgnss-qt6 ../cfg/orientable-libgnss.yaml
./orientable-libgnss-gtk4 ../cfg/orientable-libgnss.yaml

Note: The log/ and output/ directories will be automatically generated in your current working directory if they do not exist.

✨ Key Features

Zero-Allocation Signal Path: Optimized high-frequency loops (acquisition/tracking) use pre-allocated scratch buffers for maximum performance and deterministic timing.
Two-Stage Acquisition: Coarse FFT-based parallel code-phase search followed by fine frequency estimation (40 ms coherent integration with nav-bit-edge hypothesis testing) for ~6 dB sensitivity improvement.
VSM C/N₀ Estimation: Carrier-to-noise density ratio computed using the Variance Summing Method (Van Dierendonck et al.), providing front-end-independent signal quality metrics.
Carrier Lock Detection & Cycle-Slip Flagging: Normalized phase discriminator lock indicator with Doppler-predicted vs. measured carrier phase comparison for cycle-slip detection. LLI flags propagated to RINEX output.
Lock-Free Synchronization: High-concurrency buffer management using proper C11 stdatomic.h thread-safe operations to minimize thread contention.
Highly Optimized Math: Replaced FFTW3 with PFFFT for superior performance and simpler licensing, coupled with manual AVX2 and NEON SIMD intrinsics for vector math.
Hardware-Accelerated Viterbi Decoding: Includes both SSE2 and newly ported ARM NEON accelerated Viterbi(2,7) decoders for fast convolutional FEC decoding on multiple architectures.
Cache-Aligned Memory: Employs 64-byte memory alignment (via posix_memalign) to prevent false sharing and optimize multi-core cache performance.
Modular MAC/PHY Architecture: Clean separation between Physical Layer (PHY) signal processing and Media Access Control (MAC) navigation message decoding.
Broad Hardware Support: Native drivers for RTL-SDR, BladeRF, SiGe GN3S, and NSL Stereo.
Hardened Security: Bounded string operations and comprehensive error checking throughout the core engine.

🧪 Testing

The project includes a comprehensive unit test suite powered by GoogleTest.

# From the build directory
ctest --output-on-failure

For notes on receiver integration testing against SDR generators (like libgnss-tx-swift), please see the Integration Testing and Bug Fixes document.

🤝 Contributing

Please read our Contributing Guidelines to get started.

📄 License

This project uses a mixed licensing model:

Core Library & CLI: Licensed under the GNU General Public License v3.0.

See the LICENSE.md file for full details.