Sensor localization is crucial for the configuration and applications of wireless sensor network (WSN). A novel distributed localization algorithm, MDS-DC was proposed for wireless sensor network based on multidimensi...Sensor localization is crucial for the configuration and applications of wireless sensor network (WSN). A novel distributed localization algorithm, MDS-DC was proposed for wireless sensor network based on multidimensional scaling (MDS) and the shortest path distance correction. In MDS-DC, several local positioning regions with reasonable distribution were firstly constructed by an adaptive search algorithm, which ensures the mergence between the local relative maps of the adjacent local position regions and can reduce the number of common nodes in the network. Then, based on the relationships between the estimated distances and actual distances of anchors, the distance estimation vectors of sensors around anchors were corrected in each local positioning region. During the computations of the local relative coordinates, an iterative process, which is the combination of classical MDS algorithm and SMACOF algorithm, was applied. Finally, the global relative positions or absolute positions of sensors were obtained through merging the relative maps of all local positioning regions. Simulation results show that MDS-DC has better performances in positioning precision, energy efficiency and robustness to range error, which can meet the requirements of applications for sensor localization in WSN.展开更多
Three kinds of methods for processing the data of the multi-wavelength pyrometer are presented in this paper and are named curve auto-search method, curve auto-regression method and neural network method. Tbe experime...Three kinds of methods for processing the data of the multi-wavelength pyrometer are presented in this paper and are named curve auto-search method, curve auto-regression method and neural network method. Tbe experimental results indicate that the calculated temperature and the spectral emissivity compared with the true target temperature and spectral emissivity have significant deviation using the curve auto-search and the curve auto-regression methods. However, the calculated temperature and the spectral emissivity with higher accuracy can be obtained using the neural network method.展开更多
基金Supported by National Natural Science Foundation of China (No60702037)Research Fund for the Doctoral Program of Higher Education of China (No20070056129)Natural Science Foundation of Tianjin (No09JCYBJC00800)
文摘Sensor localization is crucial for the configuration and applications of wireless sensor network (WSN). A novel distributed localization algorithm, MDS-DC was proposed for wireless sensor network based on multidimensional scaling (MDS) and the shortest path distance correction. In MDS-DC, several local positioning regions with reasonable distribution were firstly constructed by an adaptive search algorithm, which ensures the mergence between the local relative maps of the adjacent local position regions and can reduce the number of common nodes in the network. Then, based on the relationships between the estimated distances and actual distances of anchors, the distance estimation vectors of sensors around anchors were corrected in each local positioning region. During the computations of the local relative coordinates, an iterative process, which is the combination of classical MDS algorithm and SMACOF algorithm, was applied. Finally, the global relative positions or absolute positions of sensors were obtained through merging the relative maps of all local positioning regions. Simulation results show that MDS-DC has better performances in positioning precision, energy efficiency and robustness to range error, which can meet the requirements of applications for sensor localization in WSN.
基金Sponsored by the National Natural Science Foundation of China(Grant No. 60377037)the Scientific Research Foundation of Harbin Institute of Technol-ogy (Grant No. HIT. 2002. 18)the Spaceflight Support Foundation.
文摘Three kinds of methods for processing the data of the multi-wavelength pyrometer are presented in this paper and are named curve auto-search method, curve auto-regression method and neural network method. Tbe experimental results indicate that the calculated temperature and the spectral emissivity compared with the true target temperature and spectral emissivity have significant deviation using the curve auto-search and the curve auto-regression methods. However, the calculated temperature and the spectral emissivity with higher accuracy can be obtained using the neural network method.
