The Study And Simulation Of WCL Localization Based On RNZ Algorithm

In this thesis, the theory of wireless sensor network (WSN) and the localization technique of nodes are introduced, including related basic principle of nodes locaion and some typical algorithms. Afterward, this paper analysed the advantages and disadvantages of these algorithms and particularly studied the a localization algorithm to calculate the distance and coordinate based on RSSI signal propagation model and centrel localization algorithm, named WCL (Weighted Central Localization) specially.Then, we find out that there are some defects within the algorithm, that is:the result of localization is always not very precise and the localization error mainly comes from obstacles. That is because the obstacles exert a tremendous influence on the signal propagating in air. It is well known that the signal is generally spread along the straight line in air, when it comes across obstacles, it will occur reflection refraction or diffraction, thereby affecting the signal strength.To the question of the influence of obstacles, this paper puts forward the improved measre. In view of the influences come from obstacles, this paper comes up with a region partition algorithm, which expresses the actual rooms to logical neighbor nodes series, in a region, there is almost no obstacles between two nodes, signal spreads along the stadia and we can use a uniform signal propagation model in an area. What's more, after unknown nodes sending positioning request, it can get preliminary localization estimate by this mean.Afterword, the core algorithm of this paper is given, called WCL nodes localization algorithm based on the RNZ algorithm. The algorithm firstly find out which branch area the unknown nodes is in, after realizing the approximately localization of unknown nodes, it realizes precise positioning of unknown nodes based on WCL algorithm.Finally, the algorithm of this article is simulatied by matlab, including part-simulation and integral-simulation, the part-simulation is used to set two parameters that influence localization result directly, that is the density and communication radius of reference nodes. The integral simulation is used to test the integral performance of the WCL algorithm that based on RNZ. Experimental results show that this algorithm can significantly reduce the localization error caused by influences of NLOS, the average positioning error of this algorithm is about 1.5m, which can satisfy the majority accuracy requirements of nodes localization in wireless sensor networks.