Research On Efficiency Optimization In The Process Of Multi-satellites Performing Area Target Surveillance Task

In the case of imaging reconnaissance satellite (IRS) observing area target, different part of the target can be observed as different elevation is chosen. This situation brings an optimization problem: how to choose the elevation of every access to guarantee the maximal efficiency of the whole process. Based on analysis of the military application requests for a space reconnaissance system and the system's working mechanism, this thesis studies the modeling and solving methods for the efficiency optimization problem of area target observe task using IRS. Finally, a simple IRS schedule tool which faces on area target was realized .The main contents and fruits of this paper are outlined as follows:First,on the basis of analysis of characteristic of IRS, this article summarize main constraints existing in scheduling of IRS. In order to deal with those extremely complex constraints, a two step method was used to solve this optimization problem. The first step, simulation pretreatment, was used to solve the complex constraint like duration constraint. Based on the first step, modeling and solving methods as the secondly step was use to com out the final result.Second, in the problem of IRS observing area target, the two intentions of maximal observing time and maximal observed area are conflicting. Thus, in order to deal with different requirements, two different models are structured on those two different intentions. In the case of the maximal observing time model, the solving problem can be break up into small problem which is find the elevation having maximal observing time of every access. Thus, a simple enumeration algorithm can be used to solve it. In the case of the maximal observed area model, the complexity of the problem was studied first, and come to the conclusion that this solve problem of this model is a NP problem.So, Genetic Algorithm (GA) and greedy algorithm is applied to solve this problem. The GA is an efficient one can find a better result when the scale of this problem is small and the greedy algorithm performs better when the scale of this problem is larger. The conclusion of some data statistical experiment also confirm that the GA do better in solving small scale problem while the greedy do better in solving large scale problem.Four, two example for two different models was provided to prove the modeling and solving method is an effective one. Last, a simple IRS schedule tool which faces on area target was realized. This tool realized connection of STK and Matlab, solve the problem of data transfer between this two software. And the user can get the final schedule result only by providing the primitive parameter of satellites and area target.