论文部分内容阅读
为了提高非视距(NLOS)环境下无线定位的准确性和可靠性,提出了一种利用数字广播信号进行移动台定位的神经网络方法.该方法利用神经网络的学习特性和逼近任意非线性函数的能力,建立到达时间(TOA)和到达时间差(TDOA)测量数据与坐标之间的映射关系.将神经网络的连接权值作为非线性动态系统的状态量进行估计,用基于扩展卡尔曼(EKF)的实时神经网络训练算法来训练多层感知器网络.由于基于EKF的训练算法给出的是连接权值的近似最小方差估计,其收敛性要优于误差反向传播(BP)算法.仿真结果表明,该算法在NLOS环境下有较高的定位精度,性能优于BP基的神经网络算法和最小二乘算法;且该定位方法不依赖于特定的NLOS误差分布,也无需视距(LOS)和非视距识别.
In order to improve the accuracy and reliability of wireless location in non-line-of-sight (NLOS) environment, this paper proposes a neural network method to locate mobile stations using digital broadcast signals. This method uses the learning characteristics of neural networks and approximates any nonlinear function (TOA) and Time Difference of Arrival (TDOA), and the coordinates are calculated.The connection weight of neural network is estimated as the state quantity of nonlinear dynamic system, and the weight of the system based on extended Kalman (EKF ) Training algorithm to train multi-layer perceptron network.Because EKF-based training algorithm gives approximate minimum variance estimation of connection weights, its convergence is better than error backpropagation (BP) algorithm. The results show that this algorithm has higher positioning accuracy and better performance than BP neural network algorithm and least squares algorithm in NLOS environment. Moreover, this method does not depend on the specific NLOS error distribution and does not need LOS ) And non-line-of-sight identification.