Distributed multi-agent SLAM for autonomous rover swarms in unknown terrain - ON-1095

Genre de projet: Recherche
Discipline(s) souhaitée(s): Études aérospatiales, Génie, Génie - informatique / électrique, Génie - mécanique
Entreprise: Anonymous
Durée du projet: Flexible
Date souhaitée de début: Dès que possible
Langue exigée: Anglais
Emplacement(s): Mississauga, ON, Canada
Nombre de postes: 1
Niveau de scolarité désiré: Études de premier cycle/baccalauréatMaîtriseDoctoratRecherche postdoctoraleNouvelle diplômée/nouveau diplômé
Ouvert aux candidatures de personnes inscrites à un établissement à l’extérieur du Canada: No

Au sujet de l’entreprise: 

We're an early-stage Canadian robotics company developing autonomous rover swarms for terrestrial exploration. Our core expertise lies at the intersection of multi-agent autonomy, SLAM, and robotic hardware integration. We aim to revolutionize surface operations in extreme and remote environments — beginning with applications in mining and defense, and scaling towards lunar exploration.

At the heart of our mission is the belief that complex, large-scale surface exploration doesn’t require complex, large-scale machines. Instead, we’re building distributed systems of small, intelligent robotic agents capable of working collaboratively to map, navigate, and adapt to dynamic terrain — autonomously and without reliance on GPS or pre-mapped environments.

By fusing cutting-edge autonomy software with scalable, rugged rover designs, we seek to become the infrastructure layer for surface robotics across industries and planetary bodies. We’re currently focused on developing our first simulation and design prototype as part of our technical milestone roadmap, with ongoing efforts to build a world-class team and secure early-stage partnerships across government, academia, and private industry.

Veuillez décrire le projet.: 

This project focuses on the development of a distributed SLAM framework for autonomous multi-agent rover swarms operating in unknown and unstructured environments. The objective is to enable a group of rovers to collaboratively explore and map terrain without prior knowledge of the environment, sharing data to construct a global map in real time.

The end goal is a robust, scalable simulation prototype where three or more virtual rovers perform coordinated surface mapping in a simulated lunar or terrestrial environment. This prototype will serve as the technical foundation for future flight-ready systems intended for lunar exploration, remote Earth applications (e.g. mining or disaster response), and commercial partnerships with space or defense organizations.

Intern tasks will include:
• Developing the core SLAM algorithms (e.g. graph-based, particle filter, or visual SLAM)
• Implementing multi-agent communication protocols for distributed mapping
• Simulating rover behavior using ROS 2 and Gazebo or Isaac Sim
• Integrating sensor models (e.g. lidar, stereo/depth cameras)
• Conducting experiments to benchmark accuracy, robustness, and system scalability

This is a research project that draws from robotics, AI, and systems engineering. Techniques will include ROS 2 development, probabilistic robotics, multi-agent systems theory, and simulation-based validation. The resulting work will support both academic publications and commercial implementation in the space sector.

Expertise ou compétences exigées: 

The ideal candidate will have a strong background in robotics, computer science, or engineering, with demonstrated experience in robot autonomy, multi-agent systems, and robot perception. A solid understanding of SLAM is essential.

Required Skills:
• Proficiency in Python and/or C++
• Experience with ROS/ROS 2 for robotic middleware
• Familiarity with robot simulation platforms (e.g., Gazebo, Isaac Sim)
• Understanding of probabilistic robotics, localization, and navigation
• Strong knowledge of reinforcement learning (RL) principles and application to robotics
• Experience implementing or adapting RL algorithms for autonomous navigation (e.g., PPO, DDPG, SAC)

Assets:
• Prior work on multi-agent coordination, swarm robotics, or decentralized decision-making
• Knowledge of graph-based SLAM libraries (e.g., GTSAM, Cartographer, hdl_graph_slam)
• Familiarity with frontier-based exploration, map merging, and autonomous planning
• Experience using ML libraries like PyTorch, TensorFlow, or Stable Baselines3 for robotic tasks
• A demonstrated passion for space exploration and space-related technologies

Strong communication and initiative are critical, especially in a fast-paced, early-stage start-up environment where autonomy and impact go hand-in-hand.