Autonomous Spacecraft Rendezvous & Docking

A simulation-first GNC project for autonomous relative navigation, guidance, control, and docking of a chaser spacecraft with a target spacecraft.

Goals

• Model relative orbital motion and spacecraft states.
• Simulate noisy relative-navigation sensors.
• Estimate relative position/velocity with a Kalman filter.
• Generate approach trajectories through hold points and a docking corridor.
• Track guidance commands with closed-loop control.
• Evaluate docking safety, accuracy, propellant use, and robustness.

Starter milestone

The initial vertical slice uses Clohessy-Wiltshire/Hill relative dynamics, a linear Kalman filter, waypoint guidance, and LQR control. Later milestones can add nonlinear orbital dynamics, EKF/UKF estimation, attitude dynamics, keep-out zones, constrained/MPC guidance, Monte Carlo testing, and hardware-in-the-loop work.

Repository layout

• src/dynamics/ — relative orbital and rigid-body dynamics
• src/navigation/ — state estimation and filtering
• src/guidance/ — rendezvous trajectory/waypoint generation
• src/control/ — feedback controllers and actuator commands
• src/sensors/ — sensor models and measurement simulation
• src/simulation/ — scenario orchestration
• src/visualization/ — plots and diagnostics
• configs/ — reproducible scenario definitions
• notebooks/ — investigations and derivations
• tests/ — unit/integration tests
• docs/ — equations, architecture, and experiment notes
• results/ — generated outputs (ignored by Git)

Quick start

python -m venv .venv
# Windows: .venv\\Scripts\\activate
# macOS/Linux: source .venv/bin/activate
pip install -e ".[dev]"
python -m src.simulation.run --config configs/baseline.yaml
pytest

Suggested roadmap

1. Baseline CW relative-motion simulation.
2. Noisy range/bearing or relative-position sensing.
3. KF/EKF relative-state estimation.
4. Waypoint/hold-point rendezvous guidance.
5. LQR/PID tracking and docking corridor logic.
6. Keep-out zones, abort logic, and actuator constraints.
7. Monte Carlo robustness studies.
8. Nonlinear orbital + 6-DOF attitude model and advanced guidance/control.

This repository is intentionally a scaffold: the architecture, interfaces, tests, and first runnable simulation are provided, while the deeper GNC algorithms are meant to be developed and documented as the project grows.