Infrastructural Positioning network

Abstract

In the last decade, Particle swarm optimization (PSO) has been applied to many applications and particularly in indoor localization which is one of the representatives among them. However, the performance of indoor localization is not satisfactory due to boundary conditions defined for initialization stage of PSO which may lead to occur premature convergence and finally an inaccurate result is obtained. In this project, I proposed to introduce clustering skill into traditional POS for creating smaller subset area before applying POS application. By using the clear relationship between distance and RSS provided by Zigbee, we can create the signal map for cluster area. After that, by using POS in this specific area, we can find a relatively precise coordinate for localization. And the testing result has proved that this new method Clustering PSO can improve the accuracy over 10% to the traditional PSO. Finally, the algorithm is applied to locate the position of robot equipped with remote control robotic arm which is expected to be applied in logistic system.