基于广播星历和神经网络的GPS卫星坐标计算方法

Calculating GPS satellite coordinates based on broadcast ephemeris and neural network method

通过在开普勒轨道模型中加入神经网络算法,得到一种全新的卫星坐标计算的混合模型:基于广播星历和神经网络混合模型,从而提高了利用广播星历计算卫星坐标的精度。计算结果表明:混合模型较传统的开普勒轨道模型在X、Y、Z3个方向上卫星坐标计算的精度提高了约10%。

A new computational model is given of the satellite coordinates generated with adding neural net- work algorithm to the Kepler track model. Based on broadcast ephemeris and neural network hybrid model, the calculation accuracy of the satellite coordinates by broadcast ephemeris will be promoted. The results show that hybrid model is more accurate than the traditional model of Kepler orbits in X, Y, Z coordinates axis by 10% of the improvement.