MoCArU

Overview of MoCArU

MoCArU is a novel Motion Capture system based on ArUco developed at Tohoku University. It is designed for localizing swarm ground robots with high precision, cost-effectiveness, and portability. It combines odometry and ArUco markers, recognized through image data processed by a remote PC, to determine the position and orientation of robots with low computational overhead. The system utilizes High Dynamic Range (HDR) cameras mounted on adjustable stands to capture images of the environment, which are then wirelessly transmitted for processing. MoCArU achieves precise localization with a root mean square error of about 0.1345 meters (around the size of the TurtleBot3 Burger model used in the evaluation experiment) using ten cameras. Its architecture allows for easy installation and movement, thanks to lightweight camera stands and wireless communication. With the ability to update camera positions dynamically and integrate robot odometry data through an Extended Kalman Filter, MoCArU is well-suited for applications requiring reliable real-time robot localization. The system's details and open-source implementation are made publicly available for broad application in various robotics domains. 

Code

We have made a GitHub repository detailing our code and instructions. You can freely find it here: https://github.com/BRN-Hub/MoCArU. 

Paper: https://ieeexplore.ieee.org/document/10394661/

Demo

The illustration is given below.