In service to the Navys stated goal to enable more complex missions with dissimilar collaborative systems, Near Earth Autonomy and its partner Carnegie Mellon University propose to leverage recent advances in miniaturization and sensing, autonomous landing and collision avoidance, and semantic mapping to increase the overall capability of unmanned systems. We will focus our attention on two operational scenariosfirst, a ground vehicle autonomously maps a landing zone, annotating it for geometric and semantic hazards such that an approaching air vehicle that is informed of the hazards can descend quickly without loitering to map the site itself. Second, we will consider the case of an air vehicle that helps a ground vehicle navigate without the use of GPS by providing a global aerial map as a reference. We will focus on designing a collaborative system, show a proof of principle demonstration, and produce and quantitative evaluation of our solution.
Benefit: The state-of-the-art in unmanned aerial systems (UAS) deployment consists essentially of vehicles operating as single entities. Upon successful completion of the proposed work, we expect to make the case that UAS will be able to effectively collaborate with dissimilar systems such as unmanned ground or surface vehicles (UGV/USV), which in turn will enable a variety of capabilities not possible today. A simple example is that of a UGV creating a detailed map of a landing site that the UAS can use to approach and land safely. A related example is that of a USV and UAS coordinating the UAS approach to and landing onto the USV. Conversely, the a UAS, with its top-down vantage point, can create maps that inform the UGV of safe areas to traverse to a meeting point where the UGV could be used to unload the UAS. UAS and other vehicles will also be able to collaborate to create high-resolution, annotated maps of large areas in support of their own mission or a subsequent troop deployment. From a system robustness point of view, multi-vehicle collaboration will enhance each platform's capability and enable more complex, higher importance missions that any one platform could not achieve individually. We also expect an additional level of reliability/redundancy from such a scheme.
Keywords: Mapping without GPS, Mapping without GPS, Architecture for Dissimilar Vehicles Collaboration, Heterogeneous Multi-Robot Collaboration, unmanned aerial system, unmanned surface vehicle, Unmanned Ground Vehicle
