Abstract:

The problem of decomposing observation tasks is considered from the customer’s point of view, where each observation task is associated with a target and the aim is to generate meta-tasks for subsequent selecting and scheduling observations. The critical component of task decomposition is calculating the access time windows and lateral swing roll angles, by means of which it is possible to observe targets by satellites. Because the polygon target can not be observed at a time, it is necessary  to segment the polygon target into some subareas before decomposing the task. This paper presents a novel task decomposition algorithm for spot targets based on a space geometry model. On this basis, combining with the dynamic segmenting method, a better task decomposition algorithm for polygon targets is proposed, and the relationship between segmenting parameter and observation efficiency is analyzed. Experimental results show that the algorithm is effective.