By providing ODyN as an open service, we aim to benefit fellow researchers and related communities while inviting feedback from external sources to enhance and expand its capabilities.
What is the project about?
In the fields of robotics and geodesy/mapping, a specific type of factor graph known as Dynamic Networks has emerged over the past decade, designed to estimate the trajectory of an agent (e.g., a robot, drone, airplane) by tightly integrating data from its on-board sensors. This is a crucial task for various applications, including autonomous navigation, simultaneous localization and mapping (SLAM), and precise photogrammetric and lidar mapping.
Open Dynamic Network, “ODyN”, offers a straightforward web-interface for trajectory estimation. Users can input data, adjust settings, visualize and export results seamlessly. Designed as an R-Shiny application with a user-friendly Graphical User Interface (GUI), it integrates the ROAMFREE sensor-fusion library and the EPFL Dynamic Network adjustment application. Processing happens on backend servers. Additionally, the related GitHub repository provides various example datasets.
Why Open?
Achieving accurate agent trajectories poses a common challenge for scientists and engineers working in robotics, navigation, and geodesy. ODyN offers a cutting-edge, adaptable solution to address this challenge effectively across various scenarios and setups. By providing ODyN as an open service, we aim to benefit fellow researchers and related communities while inviting feedback from external sources to enhance and expand its capabilities
Who benefits from it?
ODyN is valuable for researchers and engineers tackling challenges in dynamic trajectory estimation, including localization and mapping. Since its launch, it has been utilized by both practitioners, often with expertise in robotics and SLAM, and researchers in navigation and mapping. Additionally, it has been employed for educational purposes in a specialized workshop organized by the International Society for Photogrammetry & Remote Sensing (ISPRS).
How did you make it open?
The core features of the ROAMFREE sensor fusion library were first developed in C++ by Dr. Davide Cucci during his PhD. Later, during his postdoc at EPFL under the guidance of MER Jan Skaloud, these features were expanded and the Dynamic Network application was integrated. The user-friendly web interface was introduced in 2022, thanks to a collaborative effort between EPFL’s ESO group and ENAC-IT 4 research, making it freely accessible as a service. ODyN is presently undergoing active development, with enhancements being made based on user feedback and resource availability.
Contact :
