An Unscented Kalman filter based maximum likelihood ratio for NLOS bias detection in UMTS localization

In this paper, a new location tracker for cellular networks in mixed line-of-sight (LOS)/non-line-of-sight (NLOS) environments is presented. NLOS situations result in biased UMTS measurements such as Time of Arrival (TOA) or Angle of Arrival (AOA), hence in erroneous position estimates. We propose to consider NLOS as abrupt changes affecting the UMTS system which can be identified by fault detection and isolation (FDI) algorithms such as the generalized likelihood ratio (GLR) or the marginalized likelihood ratio (MLR). As the measurements depend on the mobile location in a non linear way, we present an Unscented Kalman filter based MLR to jointly identify the biased measurements and track the mobile position. Numerical results show that the developped method improves localization accuracy with a reasonable computational cost.