基于无线传感器网络定位系统中影响定位精度的非视距误差问题,通过研究最小二乘法本身的敏感性以及与最小一乘法的稳健性特点提出了组合中位最小二乘定位算法,在抑制非视距误差的同时,避免了最小一乘法计算困难的弱点,仿真证明,此算法在有少量非视距误差存在时,可以非常有效地抑制非视距误差。同时此算法与进行测距的物理层信号无关,也不需要任何的统计信息,有一定的实用价值。
In this paper,a robust wireless sensor network location algorithm,Combination Median Least Squares(CMLS) is proposed ,for the Non-Line-of- Sight(NLOS). The sensitivity of least squares multilateration about outliers and the robustness of the least absolute deviation about outliers are analyzed, and the algorithm overcomes the shortcomings of complexness of the least absolute deviation. The simulation results show the good performance of pro- posed algorithm when a few non-line-of-sights exist. Furthermore,the algorithm is independent of physical layer and do not need any prior information about the statistical properties of measurement errors or characterization of the environment where the sensor nodes are deployed. So it is a practical algorithm.