Research On Rapid Optimization Design Of Global Revisit Constellation In Low Earth Orbit

With the development of satellite technology and the reduction of cost,the global(LEO)constellation planning boom has arisen.In these constellation mission planning,the large number of satellites and the time-consuming optimization design work bring enormous challenges to the actual design work.Therefore,it is of great significance to study the rapid optimization design of constellation orbit.This paper focuses on three aspects: improving the calculation speed of covering time window model,reasonably setting sampling points of grid analysis method,and parallelizing the optimization design program,to study the fast optimization design method of constellation orbit.Specific research work is as follows:Firstly,a high precision coverage time window model of satellite sensor is established.The model takes into account the ellipsoid model,J2 perturbation and precession nutation pole shift.It is suitable for many field of view models,such as conical,rectangular,swing and so on.The hybrid quadratic polynomial is used to fit the constraint equation to quickly solve the coverage time window.When the error of time window is less than 1 second,the speed of this method is increased by more than 99% compared with the conventional recursive method.Under the same simulation conditions,the calculation accuracy of coverage time window is equivalent to that of STK by choosing appropriate interpolation node spacing.Secondly,on the basis of studying the spherical meshing model,this paper uses the regular 20-hedral meshing model to design sampling points.The spacing of sampling points is uniform,which is more efficient and reasonable for statistical constellation coverage.The simulation results show that,compared with the grid model with uniform latitude and longitude distribution,the sampling points of the grid model with regular 20-hedral partition are reduced by more than 25%,and the average errors of maximum revisit time and average coverage statistics are smaller,which is conducive to the rapid optimization design of constellations.Then,multi-core and cluster parallel computing technology is studied,and the algorithm of constellation coverage analysis module is parallelized.After the transformation,the computing speed of the program is greatly improved.The simulation results show that the computing speed on 4-core(single-processor)computing platform is more than 3 times faster than that on 24-core(6-processor)cluster computing platform is more than 18 times faster.Parallel genetic algorithm is used to optimize constellation design.When the computing time of fitness function is much longer than the communication time,the acceleration ratio of four core computing platforms is 3.2 times and that of 24 core cluster computing platforms is 22.3 times.Finally,based on the background of a microwave remote sensing observation task,the model is applied to the rapid optimization design,and various design results are given.The model can provide the optimal design results for the actual constellation design task quickly.In this paper,a fast optimization design model of global re-visiting constellation is established,which combines the methods of polynomial interpolation,fast calculation of sensor coverage time window,mesh model of icosahedral partition and multi-core/cluster parallel computation,improves the speed of optimization design.