Satellite Mission Scenario Generation And Task Attribute Intelligent Configuration Based On Gan

Satellite mission attribute configuration decision is an important premise of satellite mission planning and resource scheduling.The general process is to rationally configure satellite payload type,resolution,and task priority based on different aerospace integrated scene information such as target shape,geographic location,and meteorological conditions.The traditional satellite mission attribute configuration process usually relies on the experience and reasoning of the experts in the aerospace field.With the increasing degree of dependence on aerospace information in various fields such as national defense security and economic development,it is not necessary to manually make satellite mission attribute allocation decisions.The disadvantages of low efficiency,long time spent,and lack of human resources are increasingly prominent in the process,and it is difficult to meet the needs of future comprehensive application and service of aerospace information.In this regard,this paper combines the theory and method of Generative Adversarial Nets(GAN),and studies the satellite task scenario generation and task attribute intelligent decision-making,including the following aspects:(1)According to the actual satellite mission attribute configuration process,the satellite mission attribute configuration rule model based on the target information is established,and the basic elements such as input,output and constraint conditions of the satellite mission attribute configuration model are clarified,and the knowledge and experience of the space expert are given.The typical satellite mission attributes are configured with sample rules,and a numerical representation method of satellite task attribute configuration rules based on widedomain mode and narrow-domain mode is proposed,which provides a basis for subsequent research.(2)Aiming at the problem that the existing sample data is scarce and the configuration rules are difficult to describe accurately during the satellite task attribute configuration process,a finite sample implicit rule learning and satellite task sample expansion algorithm based on dynamic WGAN is proposed.The algorithm uses the Wasserstein Generic Grid(WGAN)as the basic theoretical framework.Through the improvement and design of the loss function,G/D network training mode and sample preprocessing method,the algorithm effectively realizes the learning of several typical satellite task attribute configuration sample rules and completes the expansion of satellite task sample.(3)Aiming at the intelligent configuration problem of satellite mission attributes,based on the large number of satellite mission sample data generated by the dynamic WGAN algorithm and conforming to the typical satellite mission attribute configuration rules,an intelligent task configuration method based on support vector machine(SVM)and WGAN network is proposed.The satellite task attribute intelligent configuration process based on the target information is realized.The validity of the WGAN extended sample library and the accuracy of the satellite task attribute configuration result are verified by simulation examples.