How to simulate GPS outages in GNSS-INS-SIM?

In the motion profile CSV, set the 'GPS visibility' column to 0 for specific command durations to mimic GPS signal loss. This allows testing algorithm robustness during periods without GNSS updates.

GNSS-INS-SIM vs MATLAB Sensor Fusion and Tracking Toolbox: which is better for research?

GNSS-INS-SIM is open-source and offers more flexibility for custom algorithm integration and sensor modeling, while MATLAB's toolbox provides a polished GUI and extensive built-in functions but at a cost. Choose GNSS-INS-SIM for full control and affordability in academic settings.

Can GNSS-INS-SIM generate data for machine learning training?

Yes, it can export synthetic IMU, GPS, and magnetometer data to files, which can be used to train ML models for navigation tasks. The configurable error models allow creating diverse datasets under various conditions.

How to add custom vibration models beyond the built-in options?

You can define a PSD model by providing a numpy array with frequency and acceleration/gyro data, as shown in the README's environment setup. For advanced models, users need to implement them manually in Python.

What's the performance like for large-scale simulations with many runs?

Performance depends on the complexity of algorithms and trajectory length; the Python-based nature might lead to slower execution for high-fidelity models or thousands of runs compared to compiled languages, but it's suitable for most research purposes.

How to visualize results in 3D for trajectory analysis?

Use the sim.plot() function with options like {'ref_pos': '3d'} to generate 3D plots of trajectories. The tool relies on Matplotlib, so customization might require additional coding for advanced visualizations.

Open-Awesome

gnss-ins-sim

MITPythonV2.2.0

An open-source GNSS/INS simulation tool that generates sensor data, runs navigation algorithms, and visualizes results for inertial navigation systems.

GitHub

1.5k stars375 forks0 contributors

What is gnss-ins-sim?

GNSS-INS-SIM is an open-source simulation tool for GNSS and inertial navigation systems. It generates reference trajectories, sensor outputs (IMU, GPS, magnetometer, odometer), and allows users to test and visualize navigation algorithms in a controlled environment. The tool solves the problem of needing physical hardware for algorithm development and validation in robotics, aerospace, and autonomous vehicle applications.

Target Audience

Navigation engineers, robotics researchers, and aerospace developers working on sensor fusion, inertial navigation, and GNSS/INS algorithm design and testing.

Value Proposition

Developers choose GNSS-INS-SIM for its comprehensive feature set, flexibility in sensor modeling and algorithm integration, and open-source nature, which eliminates the need for expensive proprietary simulation software and provides full control over the simulation environment.

Overview

Open-source GNSS + inertial navigation, sensor fusion simulator. Motion trajectory generator, sensor models, and navigation

Use Cases

Best For

Developing and testing custom GNSS/INS fusion algorithms
Generating synthetic IMU and GPS data for machine learning training
Evaluating sensor performance under different error models and vibration conditions
Comparing multiple navigation algorithms in a controlled simulation environment
Teaching inertial navigation and sensor fusion concepts in academic settings
Simulating vehicle trajectories for autonomous driving or drone navigation research

Not Ideal For

Real-time hardware-in-the-loop testing that requires direct sensor interfacing
Teams needing a GUI-based or drag-and-drop simulation environment for quick prototyping
Projects where developers lack background in inertial navigation concepts like Euler angles and Allan variance

Pros & Cons

Pros

Flexible Sensor Modeling

Supports custom IMU error definitions and built-in accuracy levels ('low-accuracy' to 'high-accuracy'), allowing for realistic sensor data generation tailored to specific hardware.

Comprehensive Algorithm Testing

Enables running multiple custom navigation algorithms in a single simulation for direct comparison, as shown in demo_multiple_algorithms.py, facilitating robust validation.

Realistic Environmental Effects

Includes vibration models (random, sinusoidal, PSD) that can be applied to IMU data, adding environmental realism to simulations without physical hardware.

Modular Algorithm Integration

Uses a standardized Python class structure for user-defined algorithms, making it easy to plug in and test custom navigation filters or fusion techniques.

Cons

Manual Configuration Overhead

Defining motion profiles requires creating CSV files with specific command types and parameters, which can be tedious and error-prone for complex trajectories.

Steep Learning Curve

Assumes familiarity with navigation concepts like reference frames (NED vs. virtual inertial) and sensor error models, limiting accessibility for newcomers without prior domain knowledge.

Limited Ecosystem and Support

As a niche tool, it lacks a broad community or extensive third-party integrations, making it harder to find pre-built algorithms or troubleshooting help compared to mainstream platforms.

Frequently Asked Questions

Related Projects

GitHub repository

Open source simulator for autonomous vehicles built on Unreal Engine / Unity, from Microsoft AI & Research

Stars18,329

Forks4,909

Last commit24 days ago

GitHub repository

Open-source simulator for autonomous driving research.

Webots Robot Simulator

Stars4,470

Forks2,032