Autonomous Robotics Systems
Research Alliance
by
Research Alliance
by
AURORA Team is proud to provide two new dataset. The Martian dataset represents a scenario designed to simulate planetary exploration, specifically focusing on the conditions encountered in Mars-like terrains. The Mine Museum dataset simulates exploration in underground environments such as lava tubes, providing a challenging context for visual navigation in low-light conditions.
This library provides a modular C++ framework dedicated on research about Visual Odometry (VO) and Visual Inertial Odometry (VIO). It also contains implementations of research works done at ISAE such as: factor graph sparsification, traversability estimation with 3D mesh and non overlapping field of view VO. This version is compatible with the middleware ROS2 (SADVio).
AURORA Team is proud to support SUPAEROMOON, a dynamic student-led initiative, in their participation in the European Rover Challenge (ERC25). This exciting project provides a unique opportunity for our students to apply their knowledge to real-world challenges in robotics and autonomous navigation.
The International conference in Automated Planning and Scheduling is very selective. ICAPS is the annual meeting of researches of the automated planning community. Déborah will present our work untitled SKATE : Successive Rank-based Task Assignment for Proactive Online Planning. SKATE can be seen as a meta-heuristic approach which successively assigns a task to the best-ranked agent until all tasks have been assigned. Our simple yet effective algorithmic approach is able to scale up to thousands of agents and tasks, performs better than the other methods in such high load conditions and even better when a variable receding-horizon is used to anticipate on the availability of agents.
AAMAS (International Conference on Autonomous Agents and Multiagent Systems) is highly recognized conference in the area of agents and multi-agent systems. Giorgio Angelotti and Nicolas Drougard will present our work on Offline Risk-sensitive RL with Partial Observability to Enhance Performance in Human-Robot Teaming. In this paper we tackled the problem of learning a POMDP model from previous experiences and selecting offline an interaction control policy optimizing a risk-sensitive metric. This policy was then used in lab experiments to control the interaction between an human operator and a robotic system in order to maximize the performace of this hetereogeneos team. Our approach presents promising results !
The topic of this paper is Active visual SLAM. To achieve robust localization and mapping accuracy, we incorporated the perception considerations in the goal selection and path planning towards the goal during an exploration mission. Through this work, we propose FIT-SLAM (Fisher Information and Traversability estimation-based Active SLAM), a new exploration method tailored for unmanned ground vehicles (UGVs) to explore 3D environments.
Krishna Murali and Elena Ponce Moreno are interested in the Mavion’s flight dynamics, i.e. the application of the laws of mechanics to the study of the drone’s trajectories, stability and control. "The aim of our research is to make the transition from vertical to horizontal flight smoother and more reliable," explains Elena. This topic, at the crossroads of aeronautics and robotics, is the subject of the paper they have submitted to ICRA and will be presenting at the conference..