Improved-PPO-Based Adaptive Safety-Critical Control for Orbit-Attitude Takeover of Failed Spacecraft With Cellular Satellites

This research investigates an adaptive safety-critical control strategy for the orbit-attitude takeover of failed spacecraft using cellular satellites under stringent motion constraints. The proposed framework, based on an improved proximal policy optimization (IPPO) algorithm, efficiently transforms the unsafe desired trajectories into safe ones, ensuring collision avoidance and protection of onboard optical payloads from direct sunlight. To address challenges such as unknown model parameters and unavailable velocity states, a radial basis function neural network (RBFNN)-based state observer is designed alongside an adaptive finite-time controller. This approach achieves high tracking accuracy without explicit parameter identification, thereby reducing the computational burden. Comparative simulations validate the effectiveness and feasibility of the proposed control scheme.