摘  要：为了克服最小二乘法在无线定位算法中的缺点，提出了基于RBF神经网络的TDOA定位算法。利用 SVM优化RBF网络权值、阈值及结点数, 获得优化稳定的RBF网络结构，将训练后的RBF网络用于TDOA定位。仿真结果表明, 该算法有很强的抗NLOS能力，与Chan算法和基于k-均值聚类法RBF神经网络的定位算法比较，具有更高的定位精度和可靠性。
关键词：无线定位；非视距传播；RBF神经网络；支持向量机

Abstract: In order to solve the problem of least-square-method in wireless location algorithm , a algorithm based on the RBF neural network was proposed. The network parameters and topology structure of RBF networks were trained in the support vector machine, Then the trained RBF networks were used to TDOA location algorithm. Simulation results show that the algorithm can mitigate the error introduced by NLOS, and the location accuracy is significantly improved and the stability of this algorithm is better than that of TDOA algorithm and the algorithm based on the RBF neural networks trained with k-mean and Chan algorithm in NLOS environment.
Key words:  wireless location; NLOS; RBF neural network; SVM