空间相机的数据处理单元在轨求解太阳耀斑位置方程时,通常采用数值迭代法进行求解,相机扫描机构根据计算结果确定指向目标的位置.由于该数值迭代方程存在一定解算时间,会使扫描机构指向耀斑目标时滞后于卫星平台传递给空间相机的瞬时轨道参数,从而带来指向位置的误差.减小该误差的一种方法就是缩短耀斑位置方程的求解时间.文章从多CPU核并行计算目标方程的角度出发,构建一种可并行化求解的数值迭代算法对耀斑方程进行求解.相比于以往的单核CPU数据处理单元,采用多核CPU结构的并行数据处理单元可以在原有的主频和硬件条件下实现计算速度的提升,从而一定程度上减小扫描机构指向耀斑目标时的位置误差.文章对耀斑位置的迭代求解方程进行了分析,进行了并行化迭代方程的构建,并在某模拟硬件计算平台上对一组耀斑观测卫星的实时轨道参数进行了仿真计算.相比于常用的单核算法,并行化耀斑位置计算方法在原有主频条件下达到了更高的精度和计算速度.%When the flare position equation is solved by space-borne processor on orbit, the usual method is numerical iterative method, due to the existence of a certain solution time, the scanning mechanism will be lag of the instantaneous orbit parameters of the flare target which was transmitted by the satellite platform, so as to bring the error of point position. One way to reduce the error is to shorten the solution time of the flare position equation. In this paper, from the point of view of the multi CPU core parallel computing objective equation, a numerical iterative algorithm for solving the flare equation is constructed. Compared to the previous single core CPU data processing unit, the multi-core CPU structure of parallel data processing unit can improve computing speed in the original frequency and hardware conditions, within a certain extent scanning mechanism can decrease the position error of flare. This paper analyzed the structure of the flare iterative solution of the equation, constructed parallel iterative equation, and calculated flare observation satellite's real-time orbit parameters in a simulated hardware computing platform. Compared with the commonly used single kernel algorithm, the parallel calculation method of the flare equation is more accurate and the calculation speed is higher than former under original dominant frequency.
展开▼
