Flow-Aware Trajectory Planning for Controlled-Altitude Balloons

Autonomous high-altitude balloons are promising tools for studying dynamic environments. Their ability to navigate strong wind currents with minimal actuation makes them efficient options for scientific research. In fact, the flow in which these vehicles operate can both facilitate and restrict their ability to traverse the environment. Thus, to deploy the balloons effectively, motion planning becomes essential, requiring the consideration of both the environmental flows and the vehicle’s ability to maneuver through them. This paper proposes a probabilistic-based kinodynamic motion planner for an underactuated variable-altitude balloon navigating in flow-dominant environments. The proposed algorithm samples feasible control inputs rather than configurations to produce thrust commands that are within the physical actuation limits of the underactuated vehicle. In this paper, we consider both two-dimensional and three-dimensional vehicles with actuation in the vertical direction only. To test our method, simulations are shown for typical environments such as the double-gyre and the jet flows, and also for a more realistic environment based on Venus’ atmospheric wind flow.