RLHAB
RL-HAB is an open-source high altitude balloon (HAB) reinforcement learning simulation environment for training automous HAB agents. HABs can leverage opposing winds to perform station keeping maneuvers for persistent area coverage of a target region over a time period of hours, days, or weeks, which can help with surveillance, in-situ stratospheric meteorologicaldata collection, or communication relays. With perfect weather forecasts, this would be a straight forward deterministic path planning problem; Unfortunately forecasts frequently have large errors in wind direction (occasionally up to 180 degrees) and also lack vertical and temporal resolution in the altitude region of interest (typically only 5-10 data points for a 10 km region), leading to significant uncertainty in flow fields.
We provide examples of training and evaluating agents with DQN in stable-baselines-3. This package also include optional integration of wandb and optuna for automated hyperparameter tuning and analysis.