Robust Trajectory Planning for Autonomous Parafoils under Wind Uncertainty

Abstract

A key challenge facing modern airborne delivery systems, such as parafoils, is the ability to accurately and consistently deliver supplies into difficult, complex terrain. Robustness is a primary concern, given that environmental wind disturbances are often highly uncertain and time-varying, coupled with under-actuated dynamics and potentially narrow drop zones. This paper presents a new on-line trajectory planning algorithm that enables a large, autonomous parafoil to robustly execute collision avoidance and precision landing on mapped terrain, even with significant wind uncertainties. This algorithm is designed to handle arbitrary initial altitudes, approach geometries, and terrain surfaces, and is robust to wind disturbances which may be highly dynamic throughout the terminal approach. Explicit, real-time wind modeling and classification is used to anticipate future disturbances, while a novel uncertainty-sampling technique ensures that robustness to possible future variation is efficiently maintained. The designed cost-to-go function enables selection of partial paths which intelligently trade off between current and reachable future states. Simulation results demonstrate that the proposed algorithm reduces the worst-case impact of wind disturbances relative to state-of-the-art approaches.

Wind Uncertainty Modeling and Robust Trajectory Planning for Autonomous Parafoils

“…Lastly, this work extends previous developments [13] to consider terminal heading constraints, by incorporating landing speed penalties into the cost-to-go function. Simulation results demonstrate that these penalties can signicantly reduce landing speed in order to encourage upwind landings, with minimal eect on accuracy.…”

“…Draper Laboratories has released 194 altitude-dependent wind proles from parafoil drops [13], collected using the sensor conguration and estimation procedure outlined in work by Carter et al [6], which are used as training data in this work.…”

“…However, as described in [13], none of these modeling approaches [1922] address the wind prediction problem over the short timescales and limited datasets inherent in parafoil precision guidance. This section ts an uncertainty model to the wind which can be incorporated into the planner to enforce robustness.…”

“…As with the conventional RRT algorithm, the parafoil CC-RRT algorithm consists of two core components: a tree growth step (Luders et al [13], Algorithm 1) that incrementally constructs the tree, and an execution step (Luders et al [13], Algorithm 2) that selects paths from the tree for parafoil execution with some frequency. Each time Algorithm 1 is called, a state is rst sampled from the environment, and the nodes nearest to this sample in terms of some heuristics are identied as candidates for tree expansion.…”

“…During the DP-means assignment step, if an observation is further than λ from the nearest cluster center, a new cluster is added with its center dened as the observation x w which yielded it. Using the aforementioned Draper wind proles [13], three distinct classes were identied.…”