Rapid trajectory time reduction for unmanned rotorcraft navigating in unknown terrain

Onboard and online flight path planning for small-scale unmanned rotorcraft requires very efficient algorithms in order to meet runtime constraints. When flying through a priori unknown environment, rapid replanning is necessary in order to maintain a safe obstacle clearance. The complexity of the planning problem rises vastly when trying to minimize flight time. This is because the helicopter's flight dynamics must be accounted for in order to produce paths which the helicopter can follow safely at high speeds. Building on the success of sampling-based planning algorithms, decoupled approaches that separate the planning of collision free and dynamically feasible flight paths into sequential steps have been shown to solve the problem very efficiently. However, these approaches may produce suboptimal results or may even fail to provide valid flight paths at all. Addressing these problems, we present an approach for planning safe, dynamically feasible and time efficient flight paths using cubic splines. Trying to reduce trajectory time, we consider the influence of path smoothing and refinement measures on the quality of the resulting flight paths and on the reliability of the planning procedure. We present comparative results in a range of benchmark scenarios, including flight through urban terrain, to evaluate the overall mission performance.