Robust indoor localization of narrowband signals

Many location-based services target users in indoor environments. Similar to the case of dense urban areas where many obstacles exist, indoor localization techniques suffer from outlying measurements caused by severe multipath propagation and non-line-of-sight (NLOS) reception. Obstructions in the signal path caused by static or mobile objects downgrade localization accuracy. We use robust multipath mitigation techniques to detect and filter out outlying measurements in indoor environments. We validate our approach using a power-based localization system with GSM. We conducted experiments without any prior knowledge of the tracked device´s radio settings or the indoor radio environment. We obtained localization errors in the range of 3m even if the sensors had NLOS links to the target device.