Resident trajectory optimization for stratospheric Airships

Resident trajectory optimization is addressed for a stratospheric Airship. The equations of motion for the airship include the factors about aerodynamic force, added mass and wind profiles are developed based on horizontal-wind model. For minimum penalty which consists of the resident error, the derivative of state variables and the energy cost during resident, a trajectory optimization problem is constructed and then converted into a dynamic programming problem by a reinforcement learning (RL) method. In different scenarios, the optimal trajectory is found by the receding horizon differential dynamic programming (RH-DDP). For the minimum resident penalty trajectory, the RH-DDP solutions which have a 57.85% and 67.3% CUP idle rate in the two cases, state that the RH-DDP method can satisfy the demands of real-time computation during resident.