Search

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.

We investigate robust cooperative localization in LOS/NLOS environments in wireless sensor networks. Round-trip time-of-arrival signal metric is considered so that time synchronization among sensors can be avoided. Owing to the non-line-of-sight effect, we model the measurement error by a two-mode Gaussian mixture distribution. However, its parameters are assumed completely unknown. We propose a centralized localization algorithm, which jointly estimates the unknown geographical coordinates and the nuisance mixture model parameters. The expectation-maximization criterion is adopted here to implement the maximum likelihood estimator. In addition, we also compute the Cramér-Rao lower bound (CRLB) for our estimation problem and present the best achievable positioning accuracy in terms of the CRLB.

We investigate robust cooperative localization in LOS/NLOS environments in wireless sensor networks. Round-trip time-of-arrival signal metric is considered so that time synchronization among sensors can be avoided. Owing to the non-line-of-sight effect, we model the measurement error by a two-mode Gaussian mixture distribution. However, its parameters are assumed completely unknown. We propose a centralized localization algorithm, which jointly estimates the unknown geographical coordinates and the nuisance mixture model parameters. The expectation-maximization criterion is adopted here to implement the maximum likelihood estimator. In addition, we also compute the Cramér-Rao lower bound (CRLB) for our estimation problem and present the best achievable positioning accuracy in terms of the CRLB.

We investigate robust cooperative localization in LOS/NLOS environments in wireless sensor networks. Round-trip time-of-arrival signal metric is considered so that time synchronization among sensors can be avoided. Owing to the non-line-of-sight effect, we model the measurement error by a two-mode Gaussian mixture distribution. However, its parameters are assumed completely unknown. We propose a centralized localization algorithm, which jointly estimates the unknown geographical coordinates and the nuisance mixture model parameters. The expectation-maximization criterion is adopted here to implement the maximum likelihood estimator. In addition, we also compute the Cramér-Rao lower bound (CRLB) for our estimation problem and present the best achievable positioning accuracy in terms of the CRLB.

We investigate robust cooperative localization in LOS/NLOS environments in wireless sensor networks. Round-trip time-of-arrival signal metric is considered so that time synchronization among sensors can be avoided. Owing to the non-line-of-sight effect, we model the measurement error by a two-mode Gaussian mixture distribution. However, its parameters are assumed completely unknown. We propose a centralized localization algorithm, which jointly estimates the unknown geographical coordinates and the nuisance mixture model parameters. The expectation-maximization criterion is adopted here to implement the maximum likelihood estimator. In addition, we also compute the Cramér-Rao lower bound (CRLB) for our estimation problem and present the best achievable positioning accuracy in terms of the CRLB.

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.

We investigate robust cooperative localization in LOS/NLOS environments in wireless sensor networks. Round-trip time-of-arrival signal metric is considered so that time synchronization among sensors can be avoided. Owing to the non-line-of-sight effect, we model the measurement error by a two-mode Gaussian mixture distribution. However, its parameters are assumed completely unknown. We propose a centralized localization algorithm, which jointly estimates the unknown geographical coordinates and the nuisance mixture model parameters. The expectation-maximization criterion is adopted here to implement the maximum likelihood estimator. In addition, we also compute the Cramér-Rao lower bound (CRLB) for our estimation problem and present the best achievable positioning accuracy in terms of the CRLB.

We study received signal strength-based cooperative localization in wireless sensor networks. We assume that the measurement noise fits a contaminated Gaussian model so as to take into account some outlier conditions. In addition, some environment-dependent parameters are assumed to be unknown. We propose an expectation-maximization based algorithm for robust centralized network localization without offline training. As benchmark for comparison, we express the best achievable localization accuracy in terms of the Cramér-Rao bound. Experimental results demonstrate the advantages of the proposed algorithm as compared to some representative algorithms.