一种基于遗传算法径向基神经网络的混合无线定位算法研究

A hybrid location algorithm based on RBF for mobile position estimation

文章提出了一种基于径向基神经网络的混合无线定位算法。采用径向基函数(Radial basis function neural net-works-RBF)神经网络建立移动台位置估计模型,并用递阶遗传算法(HGA)同时训练RBF的网络结构和参数。对进行位置估计的三种参数TOA/TDOA/AOA进行数据融合,以有效地提高定位精度。与传统BP神经网络定位算法进行比较,仿真结果表明,该算法的定位结果能够很好地满足FCC的定位要求。

In this paper, we propose an efficient hybrid location algorithm with RBF neural networks. The posi- tion estimation model is proposed with radial basis function （RBF） neural networks, and the structure and parameters of RBF neural networks are simultaneously trained by hierarchical genetic algorithm （HGA）. The model of data fusion with multi - parameters of TOA/TDOA/AOA improves hybrid location accuracy in mobile communication networks. The result shows that the algorithm with RBF neural networks satisfys the request of FCC and is more accurate than those by traditional BP neural networks.