双目标条纹搜索及GPU加速技术研究

Fast Fringe Search Algorithm for Dual Objects and GPU Acceleration

深空探测中探测器处于轨道制动期间，由于预报轨道误差较大，如果用预报轨道模型作为输入，相关处理机有时很难获得清晰的条纹，因此需要快速条纹搜索程序给出准确模型值。针对快速条纹搜索算法，研究基于图形处理器（graphicsprocessingunit，GPU）硬件平台的加速改进，同时针对后续探月任务中存在两个动态目标的情况，研究能同时计算两个目标的快速条纹搜索算法。

In deep space exploration, the relative error of spacecraft＇s predicted orbit is large during orbital maneuver. The VLBI correlator cannot obtain clear fringe if the delay models are still relying on predicted orbits. Therefore fast fringe search program is required to give high precision delay model. This thesis did research on acceleration improvement for fast fringe search algorithm based on GPU hardware. Meanwhile, for there will be two or more probes in the follow-up mission, this thesis also did research on the algorithm for two probes＇ fringe search.