Trajectory Planning of Stratosphere Airship in Wind-Cloud Environment Based on Soft Actor-Critic

The trajectory planning for stratospheric airships needs to consider the following issues: the motion characteristics of airships make the airship susceptibility to wind field disturbances; High-altitude cold clouds can change the thermodynamic equilibrium of lifting gases by altering the thermal radiation received by the airship; The wind field and cloud distribution are random and have relatively high spatiotemporal complexity. This paper proposes a trajectory planning algorithm for stratospheric airships based on Soft Actor-Critic (SAC) algorithm, a policy-based reinforcement learning algorithm. The results prove the capability of this algorithm to plan the trajectory to an arbitrary region while avoiding cold clouds and optimizing energy consumption in the time-variant wind field-cloud environment. Moreover, comparative experiments are conducted with the DQN-based algorithm, demonstrating the advantages of this algorithm in trajectory planning tasks.