| Anchor-free localization is based on wireless sensor networks. This paper analyzesthe characteristic of WSN and the feature of anchor-free localization at first. As amulti-hop, self-organizing and dynamic network, WSN is suitable for the application tothe anchor-free localization of node.In the process of anchor-free localization, it is necessary to choose the appropriatelocalization algorithm and system for different situations. In this paper, typicalanchor-free localization algorithms are studied and the multidimensional scalingprinciple is applied. The anchor-free MDS localization algorithm constructs theconnectivity diagram and relative coordinate system, which based on the connectivitybetween the nodes, and then determines the relative coordinates of the nodes accordingto the distance between the nodes. However, if the multidimensional scaling principle isdirectly applied to anchor-free localization algorithm, there may be larger errors.Therefore, there is improvement and optimization of the algorithm in this paper. Theanchor-free ordinal MDS localization algorithm is built on the monotonic decreaseconstraint of the shortest path distance of each pair of nodes, and reduces the error in theprocess of localization effectively.The anchor-free technology is suitable for combats, where the anchors are difficultto deploy, such as enemy territories and small-scale special operations, and locate thewounded on the battlefield quickly. As the background of searching and rescuing inbattlefield, anchor-free MDS localization algorithm and anchor-free ordinal MDSlocalization algorithm are applied to locate the relative coordinates of woundedrespectively. The simulation shows that the solider locations are obtained both in therectangle network and C-shaped network. Compared to anchor-free MDS localizationalgorithm, the result of anchor-free ordinal MDS localization algorithm is better. In therectangular network, the average localization error of the anchor-free ordinal MDSlocalization algorithm is reduced by1.97%, and the average localization error isreduced by4.42%in the C-shaped network. The average localization error is affected bythe communication radius. If the communication radius increases, in the rectangularnetwork, the average localization error of the anchor-free ordinal MDS localizationalgorithm is reduced by9.56%, and the average localization error is reduced by10.4%in the C-shaped network. Meanwhile, the average localization error of the anchor-free MDS localization algorithm is reduced by10.53%in the rectangular network, and theaverage localization error is reduced by12.01%in the C-shaped network. However, theaverage localization error of anchor-free MDS localization algorithm is more thananchor-free ordinal MDS localization algorithm. |
