Grid subdivision-based agile satellite scheduling method for observing area targets

Agile satellites possess roll and pitch maneuver capabilities, and adopting the strip-splicing mode for area target observation can improve the coverage rate. However, this observation process requires simultaneous optimization of roll angles, pitch angles, and the observation start and end times of each strip. To address this multistrip observation scheduling problem, an agile satellite area target observation scheduling method based on grid subdivision is proposed. The method employs hexagonal grids to perform a double-layer discretization of the target area, which is utilized for attitude mapping, coverage calculation, and heuristic guidance. With the goal of maximizing the coverage rate, a mixed-integer nonlinear programming model considering multiple constraints is established. To solve the model, a bilevel optimization framework based on adaptive feedback iteration is designed, and a bilevel optimization algorithm integrating simulated annealing with cooperative particle swarm genetic algorithm is constructed. Specifically, the upper layer performs observation strip allocation via simulated annealing and its multi-neighborhood search strategy, while the lower layer employs the cooperative particle swarm genetic algorithm to schedule pitch angles and observation windows. And the interactive iterative solution of the upper and lower layer algorithms is realized through the adaptive feedback mechanism. Extensive comparative experiments demonstrate that in agile satellite area target observation scenarios, the proposed method is superior to various comparison algorithms in terms of coverage rate and convergence stability. Moreover, compared with the traditional single-strip Earth observation mode, the strip-splicing mode of agile satellites can significantly improve the area coverage rate.