Research On Satellite Scheduling Technology For Moving Target Surveillance

Due to having wider observation area can be observed and higher security than aircraft or radar,satellite plays a very important role in various military and civilian fields,but satellite observing method has also been limited by observing width,satellite orbit and spaceborne sensor imaging performance etc.How to effectively use satellite resources for moving target search and track,which has wide motion range and strong uncertainty,has become an important topic.There are still big developments for satellite scheduling technology of moving target surveillance.Target motion state estimation as well as making surveillance plans is not only of great priority,but also a difficulty in satellite scheduling technology for moving target surveillance.On the basis of accurate prediction and estimation of moving target states,an optimal plan can be established,and the efficiency of satellite observations is closely related to the plan.Hence,this thesis studied the two aspects of satellite scheduling technology.The main research topics and innovations are as follows.1.Research on satellite scheduling strategies,a satellite scheduling algorithm combined with detection probability and information income was proposed.In order to overcome the problem caused by uncertainty from moving target and satellite inherent defects,The algorithm designed by information theory at present is likely to ignore the area with larger distribution probability in task area.An algorithm,which can adapt the searching stage and tracking stage,combined with detection probability and Kullback-Leibler(KL)discriminant was put forward for the problem.KL discriminant was introduced to measure the uncertainty,but detection probability can overcome the defect that the area which has larger distribution probability will be ignored by the KL discriminant method,meanwhile the weights can be adaptively adjusted based on real-time target distribution probability,then the satellite’s ability to find target will be improved.For the purpose of analyzing and testing the algorithm,a simulation experiment was carried out,and the computational results verify the algorithm’s validities.2.For the problem of predicting and estimating the motion state of a maneuvering target,a correction factor of model transition probability was designed based on the interactive multiple models(IMM),and estimating the motion state by particle filter(PF)optimized by Kalman filter(KF).Above all,an IMMPF algorithm whose transition probability is adaptive modified was proposed.When the motion law of the maneuvering target changed,the prediction performance of transition probability of model which is set based on prior information now will reduce.From the perspective of model matching,In order to improve the accuracy of target motion prediction,increasing the weight of current matching model in motion prediction by the correction factor of model transition probability which is designed by model probability and the change of model probability between tow surveillances.The correction factor designed above was introduced current observation result,and considering the changing trend of model matching,which is benefit to improve the accuracy of motion prediction.From the perspective of the state estimation,The traditional KF has poor effect for maneuvering target state estimation which has high nonlinear.And the calculation will increase sharply by using PF with IMM model.Therefore,the states of a maneuvering target can be estimated by PF with model information,which can control calculating amount by simplifying interactive process and reducing the increase of particles.In order to avoid the particle shortage problem,each particle was optimized by PF to improving the ability of state estimation.The simulation results indicate that the accuracy of tracking maneuvering target was improved.3.A target tracking system was designed by C# and GMAP,which integrates tracking,positioning and signal analyzing.The system mainly includes the map module,signal analyzing module,positioning modules,state prediction and estimation module.For purpose of realizing location by Time Difference Of Arrival(TDOA),simulating the working mode of spaceborne electromagnetic signal receiver of electronic satellite with four Spectrum analyzing node.The analysis and experiment which embedded target tracking algorithm proposed in this thesis verified the validities of system.Moreover,The system can be applied in future target tracking research.