Research On Improved Localization Algorithm Based On RSSI In Wireless Sensor Networks

In the application of wireless sensor networks, node location technology is a very important technology, how to know the exact location of the sensor nodes is extremely important. Normally, the area of wireless sensor networks is randomly deployed by a lot of nodes. As we are unable to know the actual location of the sensor node, at first, achieving the localization of unknown nodes is a very common technology through the transmission of information between sensor nodes. The positioning error of node localization algorithm is relatively large by the external environment, communication radius, energy consumption and so on. Therefore, how to design a smaller impact of the external environment, a higher positioning accuracy, a better positioning efficiency, a more stable and a more reliable wireless sensor network localization algorithm is a long way to go to work.This dissertation first introduces the concepts of wireless sensor networks and some classic node localization algorithm in wireless sensor networks, and analyzes the performance of this classic node localization algorithm in detail. Comparative analysis of various classical node localization algorithms from the node distribution, hardware consumption, measuring error and scalability of these four aspects, we summarize the impact of these four main factors. This dissertation carries out the further study from two aspects which are distance measuring and positioning, and improves the algorithm from these two aspects. In this dissertation, the work done on the improved algorithm is as follows:(1) Firstly, in terms of distance measuring, a Gaussian-weighted function is used to suppress the fluctuation of signal on the basis of RSSI ranging algorithm, we propose the ranging algorithm based on Gaussian-weighted correction of RSSI. The algorithm mainly uses a Gaussian-weighted correction filter to filter out the RSSI value whose error is too large and reduces the weight of RSSI value which is relatively small impact on the unknown nodes, thus obtaining higher accuracy RSSI value.(2) Secondly, in terms of positioning, on the basis of the traditional centroid localization algorithm, according to anchor nodes at different distances having different effects on the unknown nodes, we propose the distance corrected centroid localization algorithm. We reduce the affect weight of the anchor nodes at greater distances which is relatively small impact on the unknown nodes by increasing the weight of the anchor nodes at different distances, so as to improve the positioning accuracy of unknown nodes.(3) Finally, we improve the algorithm comprehensively from two aspects which are distance measuring and positioning, and then the algorithm is further optimized. Combining the distance corrected centroid localization algorithm and the RSSI ranging algorithm based.on Gaussian weighted correction, slightly improved by a method of reusing unknown nodes on the basis of the combination, we propose the centroid localization algorithm based on Gaussian-weighted correction of RSSI. This algorithm can obtain higher accuracy of RSSI value by the ranging algorithm based on Gaussian-weighted correction of RSSI. Then, we convert the RSSI value to distance between the anchor node and the unknown node, increasing the weight of each anchor node based on the distance between the anchor nodes and the unknown node. Finally, we achieve the localization of unknown nodes by weighted anchor nodes and the centroid localization algorithm. The simulation results show that the improved algorithm deletes the RSSI value whose error is too large, at the same time setting the influenced weight of each anchor node. To certain content, the improved algorithm improves the positioning coverage and accuracy. In short, the improved algorithm has higher positioning accuracy and better positioning efficiency.