FusionCore

What it does

• Estimates pose with a 23-state unscented Kalman filter that uses a declared platform profile to skip states a wheeled robot cannot observe.
• Needs only two numbers from the IMU datasheet for noise settings, with adaptive estimators handling the rest.
• Reports why each GPS fix was accepted or rejected and publishes filter health, including heading source and uncertainty.
• Includes a tool that inspects a rosbag for estimation-setup problems without requiring ground truth.
• Can convert an existing robot_localization configuration into its own format, flagging anything that does not map cleanly.
• The filter core is a plain C++ library with no ROS dependency; a separate ROS 2 wrapper is provided.

Written by Robotheca from the project's README (opens in a new tab) at commit ec593f2.

Getting started

Difficulty

Intermediate

Hardware it needs, as the project states it

Runs on a Raspberry Pi 4 in a Release build with well under 1 ms per cycle; the same source builds for ARM and x86 through Eigen. No GPU is mentioned. Targets ROS 2 Jazzy and Humble, and Docker is offered as an alternative.

How to start

The project's website (opens in a new tab)

Maintenance

Status

Active: a commit in the last 90 days

Newest commit

22 Oct 2026 on main

Latest release

0.4.0 (15 Sept 2026)

Checked on GitHub on 22 Oct 2026.

Links

• Repository on GitHub: github.com/manankharwar/fusioncore (opens in a new tab)