Multiple ground and aerial robots collaboratively explore and map, communicating and leveraging all sensing and mobility modalities across the team to maximize the coverage and efficiency by globally optimizing the plan.
Efficient hierarchical algorithmic framework enables fast autonomous exploration in 3D and complex environments where the system involves multi-modalities in sensing and mobility to map the environment.
Fast planning framework based on dynamic visibility update, can deal with both known and unknown environments where the planner attempts multiple routes and picks up environment layout along with the navigation.
Aimed at leveraging system development and robot deployment for aerial and ground-based autonomous navigation, containing simulation environments, fundamental navigation modules, and multiple visualization tools.
Deploys a fleet of ground and aerial robots in obscured underground space, collaboratively explores, maps, and searches for objects of interest. Final competition was held at Louisville Mega Cavern in Sept. 2021.
Fast Aerial Avoidance
Fast aerial maneuver and collision avoidance enabled by a trajectory library-based planner running onboard a Raspberry PI computer with constrained computing resources and taking a planning time <0.1ms.
Lightweight Aerial Autonomy
Project enables fast and reliable flights for lightweight UAVs, aimed at bringing aerial mobility to the next level flying at high speeds and close to structures, with an emphasis on the safety during aggressive maneuvers.