VaMEx3-VTB

Overview

The overarching goal of the VaMEx project is to explore and search for forms of non-terrestrial life in the Valles Marineris, using an autonomous swarm consisting of rovers, crawlers and unmanned aerial vehicles (UAVs). The goal of VaMEx3, the third phase of VaMEx, is to bring all the technologies to a TRL (technology readiness level) that allows for a analog field test in an region on Earth that is characterized by the same features and challenges as the Valles Marineris on Mars.

The VaMEx3 consortium is comprised of 14 partners, research institutes and companies (SMEs), funded by the DLR (German Space Agency). The computer graphics lab (CGVR) at University of Bremen is coordinating and managing the project.

The sub-project VaMEx3-VRN has two main goals:

• First, to investigate methods for the development of virtual twins that enable early validation and verification of both software and hardware components of an autonomous swarm. The approach here is "software-in-the-loop" and "hardware-in-the-loop", embedded in a modular software architecture of the virtual twin, simulating a realistic virtual environment, in this case the surface of Mars. At the same time, the virtual twin itself will be validated by way of a special use case, which is the terrain-relative navigation of UAVs. Throughout the VaMEx3 project, our virtual twin will be made available to all other partners in the consortium, in order to facilitate early and frequent testing of their algorithms, and in order to verify frequently the operation of the swarm as a whole.
• The second goal of sub-project VaMEx3-VRN is the development of "terrain relative navigation (TRN)" based on LiDAR, inertial measurement data, and further sensor technology. The ultimate goal is a space-ready software/hardware component, complete with a qualified space-ready computer, error safety, and a verification of the navigation performance using the virtual twin.

Virtual Twin

The virtual twin is our main contribution to the VaMEx3 project.

It is developed using Unreal Engine 5 and C++ and has two main goals:

• On the one hand, the sensor and actuator systems of the rovers, crawlers, and UAVs should be simulated as realistically as possible so that they can be used to test the algorithms that will later be used on the real swarm units. Sensors that are simulated in the virtual twin are RGB and depth cameras, LiDAR sensors, IMUs, radar, radio navigation and a sun sensor.
• On the other hand, an additional VR mode will be available to provide immersive high-fidelity visual feedback during the testing, i.e., 3D visualizations of sensor data, as well as natural interaction possibilities for one or multiple engineers or researchers. This allows them to dynamically interact with the system, collaboratively discuss the progress and behavior, and eventually help with complex decision-making and planning.

Environment

Sensor simulation

Publications

TBD