Cooperative Autonomous Tunnel Mapping

While unmanned air vehicles are starting to operate in complex urban and indoor environments, many challenges still remain. In the proposed project we will develop a team of flying robots that can explore and map environments such as networks of underground tunnels. We will address three key challenges: First, the tunnel network topology is unknown and will have to be discovered incrementally by the flying robots. Second, the robots provide imprecise state estimates and environmental representations and hence the generation of globally consistent maps will require specific attention. Third, the computing and communication resources on the flying robots are expected to be limited and hence coordination and map generation will need to be optimized to run in real time.Phase 1 has demonstrated a proof of principle system for tunnel exploration and mapping demonstrated in a simulation environment. In Phase 2, we will extend capabilities of the system as described above and will implement these in a physical team of robots that will be tested in two different environments. In Phase 3, we intend to address real needs in terms of exploration and mapping in military and in civilian search and rescue scenarios.We expect that such an effort will require a robust platform and an intuitive interface that allows for operation by non-specialists.