| Popis: |
We present an approach for protected contingency online planning in high-risk environments with strict regulatory boundary conditions. For this purpose, we integrate a deterministic online planning method into a UAM path planning framework, which operates within precomputed convex regions of obstacle-free space. As part of an offline planning phase, a database of trajectories and convex, obstacle-free regions along these trajectories are computed. During the online phase, transitions between any two preplanned trajectories inside the same region can be computed using a planning method with performance-guaranteed maneuvres. Within a given envelope of environmental disturbances, our method guarantees to contain the flight path inside its convex region, while ensuring deterministic attributes. We demonstrate the method in a scenario representing the intended application and give estimations for offline computation time on desktop-grade hardware. |
