An efficient IMM-UKF-Bias algorithm for mobile location in UMTS-FDD under NLOS conditions

In this paper, an unscented Kalman filter (UKF) in an interacting multiple models (IMM) algorithm is designed to operate in non-line-of-sight (NLoS) conditions for localisation and tracking manoeuvring mobile. The considered context is that of wideband code division multiple access (WCDMA) mode of the universal mobile telecommunications systems (UMTS) using time of arrival (TOA) estimates in the downlink (DL) scenario. The proposed algorithm combines the IMM and the UKF algorithms and takes into account possible large TOA error measurements (called bias) caused by NLoS. The latter method denoted by IMM-UKF-bias allows then tracking, not only the position and speed of the mobile, but also the bias due to NLoS, yielding an accurate location prediction algorithm for manoeuvring mobile in wireless communications.