Multi-Objective Optimization for Multi-Satellite Scheduling System

摘要

As satellite imagery plays more and more important roles in disaster relief and land monitoring, instantaneousness of obtaining more satellite images within the allotted time turns into a huge demand in Taiwan. Accordingly, the schedule-to-launch FORMOSAT-5 remote sensing satellite is planning to joint with currently on-duty FORMOSAT-2 satellite for the mission of daily imaging planning. As the owner and system operator, Taiwan's NSPO is responsible for developing multi-satellite scheduling system to meet with multiple objective requirements for image quality and delivery time. In this paper, a method of multi-objective genetic algorithm is developed to facilitate the multi-satellite imaging scheduling. We consider the mission planning for a scenario where a number of available satellites are scheduled to image hundreds of ground targets under the condition of specific payload constraints. Based on earth-satellite geometry analysis, this complicated NP-hard problem is carefully divided into some limited number of single orbit scheduling problems to simplify the satellite coordination. Multi-objective genetic algorithm is introduced to determine the suitable imaging time for each of its designated targets. It treats various imaging time epoch for each target as an individual. An optimal Pareto front is obtained by employing the specific crossover and mutation rule among the individuals. This Pareto front will provide useful and flexible decision making information to assist a mission planner to perform daily multi-remote sensing satellites mission plan. Simulation results show that the proposed algorithm improves the quality of solution and also provides good convergence speed for obtaining near-optimal solution. Performance analysis for different imaging scenarios is also discussed.