Abstract: Since existing three dimensional localization algorithms have the drawbacks of low positioning accuracy and poor stability,3D-LE-LPCCA algorithm was put forward.Firstly,locality preserving canonical correlation analysis is extended to three dimensional space and a mapping model between signal space and physical coordinate space is built.After solving the model,the adjacent node which is of unknown nodes in physical coordinate space is obtained.Secondly,our algorithm calculates the best positioning unit in the adjacent node set in constrains of degree of coplanar and vo-lume ratio.Finally,the coordinate of the unknown node is calculated through the best positioning unit.The simulation results show that 3D-LE-LPCCA algorithm has a good localization effect and improves the accuracy and stability of the three-dimensional localization algorithm,and reduces the energy consumption of the node.

Key words: Wireless sensor networks,Locality preserving canonical correlation analysis,Three dimensional localization algorithms,Degree of coplanar,Volume ratio