Analysis And Research On Wireless Sensor Network Localization

Wireless sensor network (WSN) , which integrates technologies of sensor, embedded system, distributed information processing and wireless communication etc, leads to a new pattern of computer network. It can be widely used in military, civil and other commercial applications and is becoming a hot topic and a promising field in wireless communication researches. The fundamental issues of wireless sensor networks are topology control, data confusion, localization, networking protocols and security. Localization is one of the key problems in WSN, because the location of node stamped in the message is crucial information without whih the message will make no sense. Some communication protocols of WSN are built on the knowledge of the geographic positions of nodes. So it is very important to determine the positions of the nodes.A brief introduction of WSN is given to describe the framework and characteristics of WSN, and the issue of sensor localization is put forward. Then, the requirements and developments of localization algorithms are explained particularly. Localization algorithms can be divided into two types: range-based methods and range-free methods. In this paper, a brief survey of these methods is given and range-based methods are stressed. In range based methods, a crucial problem is non line-of-sight (NLOS) propagation which can severely damage the precision of the location estimation. Considering the characteristics of WSN, the traits of the NLOS error and the typical NLOS error mitigation/elimination algorithms are analyzed and the unsuitabilities of these algorithms are discussed if they were used in the localization in WSN.With the consideration of the characteristics of WSN and its application environments, weighing the energy consumption of nodes, amount of communication, computation and localization accuracy, based on the residual weighting algorithm(Rwgh), two novel TOA based residual weighting algorithms are presented: Lower-Computational-Cost Residual weighting algorithm (LCC-Rwgh) and Optimized Residual Weighting algorithm (ORwgh). By picking out the subsets of range measurements with minimum mean Residual Square in different ways and calculating the weighted mean of their position estimations, these algorithms can optimize the combinations of range measurements without knowing the channel characters, it also avoided iterative communication. ORwgh's computational complexity is lower than that of the Rwgh algorithm, and LCC-Rwgh has a lower computational complexity compared to ORwgh.The performances of the algorithms are tested by simulations via Matlab7 in both large scale outdoor and small scale indoor environment according to the possible WSN application environment. Simulation results show that ORwgh has almost the same performance as the Rwgh which can mitigate NLOS error efficiently while LCC-Rwgh is suitable for the NLOS error mitigation in indoor environment and improves the localization accuracy in this environment.