基于遗传算法的RBF神经网络在GPS高程拟合中的应用

Application of RBF neural network to GPS elevation fitting based on genetic algorithm

本文针对RBF神经网络中隐含层径向基中心值的确定,利用遗传算法对其进行优化,并应用于高程拟合的实验研究中。通过将遗传算法优化的RBF神经网络与K-均值优化的RBF神经网络及标准RBF神经网络进行高程拟合的误差对比分析表明:遗传算法优化的RBF神经网络提高了拟合的稳定度,改善了精度。

The paper applied genetic algorithm（GA） to optimize the value of radial basis function centers in the hidden layer of RBF neural network. The method was used in experimental studies of elevation fitting. By comparing and analyzing the accuracy of GPS elevation fitting by GA-RBF, RBF combining K-means clustering algorithm and Standard RBF, it concluded that GA-RBF neural net- work could enhance the fitting stability and improve the accuracy.