Abstract: Concerning Chan location algorithm's poor performance in Non-Line-of-Sight (NLOS) environment, a Chan location algorithm based on the Elman neural network was proposed. The Elman neural network was adopted to correct the NLOS errors with its dynamical characteristics and non-linear approach capacity and then the position was calculated by Chan's algorithm. The simulation results indicate that the location algorithm can improve the performance even in serious NLOS environment. The performance of the proposed algorithm outperforms that of Chan's algorithm and Taylor's algorithm.

Key words: Non-Line-Of-Sight (NLOS), Elman neural network, Chan location algorithm, Time-Different-of-Arrival (TDOA)

摘要： 针对Chan定位算法在非视距(NLOS)环境下定位性能差的缺点，提出一种基于Elman神经网络的Chan定位算法，利用Elman神经网络的动态递归特性以及强大的非线性映射逼近能力，对NLOS误差进行修正，再利用Chan算法定位。仿真结果表明，在NLOS误差较大的环境下该算法仍具有良好的定位精度，性能优于Chan算法和泰勒级数展开法。

关键词: 非视距传播, Elman神经网络, Chan定位算法, 时间到达差