The Open Informatics graduate Jiří Ulrich got into the finals of the prestigious IT SPY 2022 competition with his thesis Fiducial marker-based multiple camera localisation system. It introduces a multi-camera localization system that, based on an improved algorithm, detects the black and white WhyCode circular reference marks, determines their position and orientation in space and identifies them. Built on cheap and widely available cameras, the WhyCode system represents an accessible localization system as an alternative to the current expensive and closed systems with specialized hardware.
In order to assess the performance, Jiří also tested the original single-camera system used in the field of mobile and swarm robotics. For this purpose, he created a simulated environment allowing for dynamically changing scenarios. Together with the team, he also collected a data set capturing the deployment of a system for tracking the position of a mobile robot during autonomous navigation in exteriors. The results of these experiments show higher accuracy and reliability when using the presented system compared to its single-camera form.
Alternatives of the presented system are used for determining the position and orientation of robotic swarms by a number of European universities researching swarm robotics or for monitoring the activities of queen bees within the European RoboRoyale project.
What is the purpose of the system you created?
The system is used for accurate tracking and determining the position of fast-moving objects. Current systems, e.g. PTI Phoenix or OptiTrack, need special tags for this, which they detect with expensive infrared cameras. These systems are very expensive, require long and complex preparation and maintenance to use, and have difficulty functioning in direct sunlight.
What is the biggest advantage of your system?
It's cheaper and easier because you only need a camera and a printer. No special equipment is required and if the markers get damaged, just print or draw new ones. Our system can be set up in minutes, works in direct sunlight and has a longer range than similar methods. The system is also available for the Android platform, so you just need to download a small application to test it. A big advantage is that the code is open source, so anyone can customize it as they want.
What are some potential applications of this technology?
The technology can be used in various applications for the localization and creation of formations of autonomous drones, augmented reality, determining the position of robots during interactions with each other or with the environment, or simply at any time when it is necessary to know the position of the object to which the tag is attached. Various alternatives of the system have been and are used by a number of world universities. The University of Pennsylvania uses our tags for research into drones and smart materials, the Reykjavik University created a modified version for locating marine drones, the University of Lincoln uses tags to locate mobile robots, the University of Graz tracks the movement of social insects and the University of Manchester uses markers for locating and evaluating the activity of swarm robots.
We also collaborate with several technology companies (especially their R&D departments). Synergy Logistics has been testing the application of tags in an autonomous logistics warehouse, SAGA is using tags in robotic agriculture, Zuri is investigating the use of tags in a VTOL landing maneuver, and Enchanted Tools is trying to precisely interact with humanoid robots. Furthermore, this technology can be found in the European research projects STRANDS, HAZCEPT and in the current project RoboRoyale, which combines microrobotics, biology, ethnology and chronorobotics to stabilize and support endangered ecosystems.
How long have you been working on this topic?
The WhyCode system has been a topic of mine for some time. I dealt with it already in my bachelor's thesis, where I focused on reducing the mark localization error while increasing the stability of identification when using a system with only one camera.
And what are your future plans?
The project is part of my doctoral research and a good stepping stone, because there is a significant publication interest in the topic. A big advantage of WhyCode is its diverse applications. What I really enjoy about the research is that it combines robotics and computer vision.
You can vote for Jiří in the IT SPY competition – Public Award until November 22!