Abstract-Emerging cellular networks integrate the user terminal geo-localization function besides the communication function. The conventional positioning approach is to estimate the terminal location in two-steps: first the distance to all connected base stations is assessed based on signal time-of-flight measurements, then the location is deduced from the distances by multilateration. The two-step approach incurs a performance degradation because information is lost from the received signal when the multi-lateration is performed. In this paper, we propose to iterate between the two conventional steps to progressively refine the distance estimates based on the knowledge of the position estimate obtained from the previous iterations. The information exchanged between the two-steps not only consists in the mean of the estimates (distance or position) but also of their variance that convey information about the reliability of the estimates. Simulation results show that the achievable performance after a few iterations is close to the performance of the optimal approach that directly estimates the position based on the observation of the received signal.

Recently, iterative localization has arisen as a promising approach to localize a Mobile Station (MS) in a cellular system. The conventional geo-location is obtained in a two-step approach: propagation delays are estimated and then the multi-lateration is responsible for the determination of the user position, based on the estimated delays. Iterative localization iterates between the two conventional steps to progressively refine delay estimates based on the position estimate available from the previous iterations. This localization scheme was seen to provide appealing performances compared to the two-step approach. It also seems to be computationally attractive with respect to direct localization that estimates the position using the digitized received signals directly. However, the iterative localization solution developed in literature relies on a strict time synchronization between MS and Base Stations (BSs). Moreover, the computational complexity of the iterative approach is not thoroughly compared to two-step and optimal solutions. This paper therefore proposes a new iterative localization method able to operate in a cellular system with time-misaligned terminals. We show by means of a detailed complexity analysis that the iterative positioning algorithm is one order of magnitude less complex than direct localization. Simulation results prove that the achievable performance after a few iterations approaches the performance of the direct localization solution.

No abstract

Iterative localization is arising as a promising solution to determine the position of a mobile station in a cellular network. We recently showed that in a perfect lineof-sight environment, iterating between the conventional delay estimation and multi-lateration steps allows to approach the performance of the direct localization based on the observation of the received signals. In this paper we extend our iterative localization method to operate in rich multipath environments. Simulation results prove that given some prior knowledge on the power delay profile of the channel, the proposed iterative algorithm is robust to harsh propagation environments and performs very close to the direct localization approach.

In practice, the subspace-based algorithms such as Multiple Signal Classification (MUSIC) suffer from sensitivity to antenna-array response errors and therefore they require the assessment of the calibration gain and phase perturbations. This paper evaluates experimentally the accuracy of Angle-of-Arrival (AoA) estimation based on the MUSIC algorithm only coming from these perturbations in the context of Internet-of-Thing (IoT) applications. First of all, a new Over-the-Air (OTA) calibration method is proposed and gain and phase uncertainties are investigated. The impact of these uncertainties on the accuracy of AoA estimation is then studied and compared with the theoretical analysis. The experimental results show that the calibration errors coming from hardware imperfections can cause some degrees of uncertainty in AoA estimation.