Virtual Reality Lab for Rover Navigation using Mars datasets

This paper introduces a Virtual Reality Lab, based on real information, designed to enhance the training and navigation capabilities of planetary rovers using Mars datasets. The exploration of Mars is gaining more and more interest during the last few years, thanks to the use rovers to investigate the planet surface. Efficient rover navigation is crucial for achieving successful mission objectives while ensuring the safety of the equipment. Our Virtual Reality Lab goes beyond the state of the art by including two main features that allow to recreate a real scenario from Mars. The first one is the use of AI-based super-resolution methods to increase the quality of orbital imagery from Mars; and the second one is related to the use of Generative Adversarial Network to reconstruct 3D terrain features from single images. The primary objectives of the virtual reality lab are twofold: first, to facilitate the training of rover operators in a realistic and immersive Martian environment, and second, to enable mission planning and real-time rover control through an intuitive and dynamic interface. To this end, a rover prototype was included into the virtual lab, using the software-in-the-loop concept to emulate its navigation capability.