基于遗传算法和模糊神经网络的矿区GPS高程转换

GPS Height Conversion in Mining Area Based on Genetic Algorithm and Fuzzy Neural Network

基于模糊神经网络和遗传算法的原理,确定了矿区GPS高程转换的遗传算法-模糊神经网络模型,研究了遗传-模糊神经网络、二次曲面拟合和BP神经网络3种模型在GPS高程转换中的应用。结果表明:遗传-模糊神经网络模型拟合数据的精度更高、网络性能更稳定,有效的避免了局部极值的问题,可以用于GPS高程转换。

Based on the theory of Fuzzy Neural Network and Genetic Algorithm, it determinates the geneuc algorithm, fuzzy-neural network model （GA-FNN） of GPS height conversion. And investigation on the three kinds of application model of GA-FNN, conicoid fitting and BP neural network into GPS height conversion are carried out. The test resuits show that GA-FNN has better fitting precision being higher and net function being more stability, and it can effectively avoid the problem of local extremum. So the model of GA-FNN can be used for GPS height conversion.