A Range-free Integrated Localization Algorithm In Wireless Sensor Networks

As the rapid development of Wireless Sensor Networks (WSN), it is applied extensively in many fields, such as military, disaster predication, medical, environmental monitor, academic research and etc. There are several research directions about WSN, they could be summarized as Localization, Routing, Security, and energy. Localization is the supportable technology of WSN, compared with the traditional wireless networks, such as ad-hoc, the positions of sensor nodes in the WSN are usually fixed, most applications are based on this. For instance, if fire danger class of a forest is wanted, people could broadcast sensor nodes by plane, these nodes collect related information and get it back to the monitor center. Only if the positions of sensor nodes are known, analysts could judge the position in which might have a fire, or even they predict there will be a fire in the forest, they do not know the concrete site. Consequently, localization technology is under observation by researchers all over the world, they have presented many typical algorithms in recent years. These algorithms could be classified according to different standards, and the most used standard is Range relevance, there are Range-based and Range-free localization algorithms. Range-free algorithms have many advantages, such as less requirements of hardware, little influence by application environment, avoiding the error of range measurement, so they obtain more observations.Several typical range-free algorithms have their own advantages and limitations, for example, the localization accuracy of Centroid algorithm is general, but it could reduce communicational overhead and computational overhead of sensor node, be at ease of implementation; DV-Hop algorithm will spend more communicational overhead and computational overhead, by its localization accuracy is much better. In application, we should consider to combine different algorithms which have their own features, subalgorithm is selected according to different conditions in order to express its advantages and make up for the deficiencies. Consequently, this paper presents a comprehensive algorithm called CDLI algorithm which combines Centroid and DV-Hop algorithms. It could decide to select which algorithm according to parameters, we choose Centroid and DV-Hop as subalgorithms because Centroid could reduce the communication and computational overhead of anchor nodes, and DV-Hop could improve the localization accuracy. As the anchor nodes are very expensive in WSN, this paper pays attention to the energy consumption of anchor nodes. Communicational overhead is the main part of energy consumption, CDLI algorithm could reduce the communicational overhead of anchor nodes, and save energy.CDLI brings in a parameter k which gets different values as the network environment changes, experiments show that k value is related to average connectivity of the network. CDLI selects subalgorithm according to k value. It will select Centroid to localize an unknown node if the node has no less than k anchor nodes in one-hop scope around itself; or it will select DV-Hop to localize this unknown node. In the first stage, anchor nodes broadcast packets which contain their own location to neighbor nodes, Centroid and DV-Hop all need the information. CDLI algorithm shows its difference in the second stage, the unknown nodes which select DV-Hop need to send acknowledgment messages to each anchor node to certify that they will use DV-Hop algorithm, they need related information from anchor nodes. When the next-hop of a node is an anchor, this node collects the messages from the unknown nodes, and forwards these messages to the next-hop anchor node with one packet. The anchor node realizes which of all known nodes select DV-Hop localization according to the messages forwarded by different nodes, and multicasts packets which contain average distance per hop information to those unknown nodes. Compared with the broadcasting in the second stage of the traditional DV-Hop algorithm, anchor nodes could send less packets, it will reduce the communicational overhead and save energy for anchor nodes. Although anchor nodes need to spend communicational overhead for receiving messages from the last-hop nodes, the overhead for receiving is much less than sending, and anchors do not send packets purposely which contain average distance per hop information to the unknown nodes which select Centroid, these will reduce the communicational overhead of anchor nodes. In order to prove CDLI algorithm, this paper tests it in MATLAB7.0. The experiments show CDLI improves the localization accuracy than DV-Hop by 4.5%-7%. Some of the unknown nodes select Centroid localization, the communicational overhead of anchor nodes is reduced.CDLI algorithm not only improves the localization accuracy, but also reduces the energy consumption of anchor nodes in WSN. CDLI combines subalgorithms, this idea might be referenced in designing new localization algorithms. CDLI brings the view of energy into the localization algorithm, as the anchor nodes are especially expensive in WSN, CDLI takes the energy consumption of anchor nodes into careful consideration. I hope this algorithm will show its advantages in applications. In recent years, wireless sensor networks are applied in many fields, researchers pay more attention to the localization technology, they present some algorithms adapt to different circumstances, and make improvement on them. One main research direction in the future is localization in complicated circumstances. The development of localization technology brings new space for researchers with the challenges at the same time.