New advances in control theory are required to enable aggressive maneuvering of autonomous vehicles, while adapting in real time to changes in the operational environment. A hybrid control architecture, the states of which represent feasible trajectory primitives, is constructed to reduce the complexity of the motion-planning problem for a nonlinear, high-dimensional system such as an aerial vehicle. Any feasible trajectories in the primitive list are available to the automatic control system; these may include a complete set of transitions between pairs of trim trajectories in addition to pilot-inspired behaviors recorded during manual flight tests with a human pilot. This paper describes the structure of a hybrid automaton that solves a time-optimal motion-planning problem by sequencing maneuvers in real time from such a primitive list. The algorithm can be used in a free workspace, or in the presence of fixed or moving obstacles. We present simulation results showing the effectiveness of this approach for a behavior library generated by a combination of analysis and live flight tests with a small remote-controlled helicopter.
|Original language||English (US)|
|Title of host publication||AIAA/IEEE Digital Avionics Systems Conference - Proceedings|
|Publisher||IEEEPiscataway, NJ, United States|
|State||Published - Dec 1 2000|