||For various kinds of applications in sensor network, the positional information of each sensor tag is essential.
In order to know the exact position of each tag, the location estimation method using TDOA (Time Difference of Arrival) measurements is often considered. Some conventional linear search algorithms used for TDOA-based estimation, however, are easily degraded due to serious multipath fading and Non-Line-of-Sight (NLOS) propagation especially in indoor or urban environments. In this paper, we describe some novel algorithms for accurate indoor positioning of a sensor tag in a network of ﬁxed nodes. The proposed technique uses layered particle ﬁlter. Firstly,
multiple ﬁxed nodes receive the noisy signal transmitted from the tag of interest. The adverse effects of multipath or NLOS propagation in TDOA measurements are reduced by ﬁrst particle ﬁlter. Then, the modiﬁed measurements are incorporated into the positioning algorithm by second particle ﬁlter. The non-Gaussian state space model used in the ﬁrst layer is helpful to ﬁnd outliers and compensate such adverse effects. This procedure directly inﬂuences the performance of second one. Some simulation results show the usefulness of the proposed technique which can achieve more robust and precise estimation compared with conventional systems.