Research On Efficient Range-free Localization Algorithms In Wireless Sensor Networks

A wireless sensor network (WSN) usually consists of a large number of inexpensive sensor nodes with the ability of sensing, computing and wireless communicating. In many WSN applications, such as environmental monitoring, disaster rescue and geographic-base routing, the location information of sensor nodes is crucial. Owing to high cost of Global Positioning System (GPS) receivers, only a small portion of sensor nodes are equipped with GPS receivers and these nodes are called anchor nodes. The remaining nodes, called unknown nodes, are location-unaware initially, and need to be localized with the assistance of the anchor nodes. Therefore, a localization problem, which solves the location information of location-unaware nodes, is a very fundamental and essential issue in many WSN applications. Existing localization algorithms can be totally classified into two categories: the range-based algorithms and range-free algorithms. The range-based algorithm can achieve higher localization accuracy, especially in open scenarios; however, additional hardware is required. The range-free algorithm does not need additional hardware, and, thus, it is a cost-effective approach for the WSN localization problem.In this paper, we aim to enhance the performance of the range-free algorithms. If more useful information is available, the localization accuracy of the existing range-free localization algorithms should be increased. However, it is not reasonable that the information is obtained by equipping sensor nodes with additional hardware, because the cost would be very high, which is just the reason why we choose range-free algorithm rather than range-based algorithm to solve the WSN localization problem. Therefore, how to exploit to utilize the information which is available and underutilized in present hardware to enhance the performance of the range-free algorithms is our main focus. Moreover, due to the energy limited of sensor nodes, we also concern how to decrease the computation cost of the localization algorithm.The contents of our work are listed as follows:Firstly, the conventional localization algorithms use the neighboring anchor nodes or all anchor nodes in the WSN to localize the unknown nodes. The localization accuracy of the former algorithms is coarse because of the limited information provided by the neighboring anchor nodes. The communication cost of the latter algorithms to collect the information of all the anchor nodes is very high, and they suffer from irregular network topology. Consequently, a tradeoff is that collecting the information of the anchor nodes in two hops.Secondly, many conventional localization algorithms do not consider the influence of the radio irregularity in reality. Although some researchers mention the influence, further specialized improvement of their algorithms is not made. In our work, we took radio irregularity into consideration and proposed an algorithm which is more robust against radio irregularity.Thirdly, the Received Signal Strength (RSS), which can be easily measured by present hardware, is underutilized before by range-free algorithms. In our work, we investigated the relationship between the RSS and the node-to-node distance, and then exploited to utilize it in a more elaborate way that the localization accuracy of the conventional algorithm is increased.