Fostering Research on Mobile Robotics with High-Quality Data and Open Tooling

Mobile ground robots have become increasingly popular in academia and various industrial applications. However, unlike other domains like aerial robotics, autonomous driving, and construction, there is currently no high-quality, large-scale dataset or reliable benchmark established in this field, nor the tooling available to do so. Creating such a dataset would be immensely valuable for researchers and developers in fostering research on robust and practical algorithms across diverse environments. Moreover, the development of a standardized benchmarking platform would promote fair comparisons between different approaches, fostering innovation and facilitating the rapid progress of mobile ground robot research. Motivated by this, we propose to collect and share a high-quality, versatile, large-scale robotic dataset, “GrandTour”, with scalable and automated tooling– focusing on legged robots in addition to a set of benchmarks and the necessary tooling.