strapdown-rs — Strapdown inertial navigation in Rust

Repository · Documentation · API docs · crates.io

strapdown-rs is a strapdown inertial navigation system implementation in Rust, built for research, teaching, and integration into real systems. It targets the low size, weight, and power domain — phones, drones, small UAVs and UUVs — where MEMS-grade sensors and their error characteristics dominate the navigation problem.

The project is organized as four crates:

• strapdown-core — the library: a WGS84 Earth model, the standard nine-state strapdown forward mechanization, and full-state navigation filters — an unscented Kalman filter and a particle filter — for estimating position, velocity, and attitude from IMU data.
• strapdown-sim — a command-line INS simulator. Runs open-loop dead reckoning or closed-loop filtering, and simulates the GNSS conditions that actually matter: intermittent availability, duty cycling, correlated-noise degradation, slow bias, and spoofing.
• strapdown-data — smartphone-collected MEMS IMU and GNSS data.
• strapdown-geonav — geophysical anomaly aiding (gravity and magnetic), built behind a feature flag. This is the active research edge.

Design priorities are correctness, numerical stability, and extensibility — the base mechanization and filters are meant to be a foundation you build aiding algorithms on top of, not a black box. It is deliberately not a full-featured INS: there is no raw IMU or GNSS signal processing, and the integration is loosely coupled.

cargo add strapdown-rs        # library
cargo install strapdown-sim   # simulator

Planned: an error-state formulation, plus C and Python bindings.

The accompanying software paper is under review at JOSS.