论文部分内容阅读
为了提高局部天区星图识别的实时性和抗噪声能力,提出基于FPGA平台的SOFM神经网络星图识别方法。首先从局部天区内选取导航星、构建导航三角形,然后利用SOFM网络的聚类功能对导航星三角进行归类,最后将训练好的网络应用到FPGA上,充分发挥SOFM并行性算法的优势,以并行流水线的方式来星图识别。仿真结果显示:SOFM算法比传统的三角形算法具有更强的抗噪能力,而使用FPGA进行识别比使用串行处理器在速度上可以提高20多倍。
In order to improve the real-time performance and anti-noise ability of local sky map recognition, a novel SOFM neural network map recognition method based on FPGA platform is proposed. First select the navigation star from the local area to construct the navigation triangle, then use the clustering function of SOFM network to classify the navigation star triangle, and finally apply the trained network to the FPGA to give full play to the advantages of the SOFM parallel algorithm, To parallel pipeline approach to the star map recognition. The simulation results show that the SOFM algorithm has stronger anti-noise ability than the traditional triangular algorithm, and the use of FPGA to identify than the use of serial processors can increase the speed of more than 20 times.