arXiv subjects

Xiao-Wen Chang

Publications and source records attributed to Xiao-Wen Chang.

58 records · Page 4Linked to original sources

arXiv2014

Mobile Localization in Non-Line-of-Sight Using Constrained Square-Root Unscented Kalman Filter

Localization and tracking of a mobile node (MN) in non-line-of-sight (NLOS) scenarios, based on time of arrival (TOA) measurements, is considered in this work. To this end, we develop a constrained form of square root unscented Kalman filter (SRUKF), where the sigma points of the unscented transformation are projected onto the feasible region by solving constrained optimization problems. The feasible region is the intersection of several discs formed by the NLOS measurements. We show how we can reduce the size of the optimization problem and formulate it as a convex quadratically constrained quadratic program (QCQP), which depends on the Cholesky factor of the \textit{a posteriori} error covariance matrix of SRUKF. As a result of these modifications, the proposed constrained SRUKF (CSRUKF) is more efficient and has better numerical stability compared to the constrained UKF. Through simulations, we also show that the CSRUKF achieves a smaller localization error compared to other techniques and that its performance is robust under different NLOS conditions.

stat.AP↗

arXiv2014

Distributed Cooperative Localization in Wireless Sensor Networks without NLOS Identification

In this paper, a 2-stage robust distributed algorithm is proposed for cooperative sensor network localization using time of arrival (TOA) data without identification of non-line of sight (NLOS) links. In the first stage, to overcome the effect of outliers, a convex relaxation of the Huber loss function is applied so that by using iterative optimization techniques, good estimates of the true sensor locations can be obtained. In the second stage, the original (non-relaxed) Huber cost function is further optimized to obtain refined location estimates based on those obtained in the first stage. In both stages, a simple gradient descent technique is used to carry out the optimization. Through simulations and real data analysis, it is shown that the proposed convex relaxation generally achieves a lower root mean squared error (RMSE) compared to other convex relaxation techniques in the literature. Also by doing the second stage, the position estimates are improved and we can achieve an RMSE close to that of the other distributed algorithms which know \textit{a priori} which links are in NLOS.

cs.NI↗

arXiv2012

Partial LLL Reduction

The Lenstra-Lenstra-Lovasz (LLL) reduction has wide applications in digital communications. It can greatly improve the speed of the sphere decoding (SD) algorithms for solving an integer least squares (ILS) problem and the performance of the Babai integer point, a suboptimal solution to the ILS problem. Recently Ling and Howgrave-Graham proposed the so-called effective LLL (ELLL) reduction. It has less computational complexity than LLL, while it has the same effect on the performance of the Babai integer point as LLL. In this paper we propose a partial LLL (PLLL) reduction. PLLL avoids the numerical stability problem with ELLL, which may result in very poor performance of the Babai integer point. Furthermore, numerical simulations indicated that it is faster than ELLL. We also show that in theory PLLL and ELLL have the same effect on the search speed of a typical SD algorithm as LLL.

cs.IT↗

arXiv2012

Column Reordering for Box-Constrained Integer Least Squares Problems

The box-constrained integer least squares problem (BILS) arises in MIMO wireless communications applications. Typically a sphere decoding algorithm (a tree search algorithm) is used to solve the problem. In order to make the search algorithm more efficient, the columns of the channel matrix in the BILS problem have to be reordered. To our knowledge, there are currently two algorithms for column reordering that provide the best known results. Both use all available information, but they were derived respectively from geometric and algebraic points of view and look different. In this paper we modify one to make it more computationally efficient and easier to comprehend. Then we prove the modified one and the other actually give the same column reordering in theory. Finally we propose a new mathematically equivalent algorithm, which is more computationally efficient and is still easy to understand.

cs.IT↗