MWLS: median-weighted least squares algorithm for in-door localization in LOS environment

The weighted least squares (WLS) algorithm is widely used in localizing under indoor environment due to its advantage of simple calculation. However, its accuracy is not well because it is sensitive to noise and it ignored the effect of observations. Therefore, this paper proposes a median weighted least squares algorithm (MWLS), which combines median observations (MO) and a new weighted matrix, namely, the expectation of the difference between each observation and MO to improve positioning performance. The simulation results show that the MWLS algorithm is superior to WLS in both accuracy and stability.