Research On Node Selection Algorithm For Cooperative Localization

With the rapid development of mobile internet technology,a variety of services and applications based on wireless positioning have been rapidly popularized and promoted.Indoor positioning is one of the most important applications of next-generation networks.Compared with traditional indoor positioning,multi-terminal cooperative positioning solves the problem that the wireless positioning is low accuracy under the condition that the number of the visible reference nodes is insufficient.It is different from manually placed reference nodes with precise location information.Terminal nodes as the reference nodes often have errors.What's worse,when Terminal nodes are intensive distribution,the interference of data transmission will be exacerbated,and the ranging error will be increased.The extensive cooperative of heterogeneous networks and the interconnection of terminal nodes are more likely to lead to large-scale diffusion and non-linear accumulation of errors.Aiming at the problems existing in cooperative localization,the theory of multi-terminal cooperative localization is studied in this thesis,and a series of high performance algorithms are proposed.The detailed work is as follows:Firstly,the concept of factor graph is introduced into the probabilistic cooperative localization algorithm,and the inter-node factor graph and the internal factor graph framework are constructed by using the probability information in the wireless parameter measurement.Then the process that the belief information iteratively propagated on the factor graph is analyzed.Based on this model the thesis introduces the SPAWN algorithm.Then,aiming at the problem of cooperative localization in NLOS environment,on the premise of guaranteeing network coverage and performance,this thesis proposes a cooperative node selection algorithm based on Fisher information matrix.This algorithm can make the nodes that to be located select partial nodes from all reference nodes to take part in cooperation.The selection metric used in the algorithm can evaluate the influence of reference nodes on the accuracy of position estimation of terminal node.Thus,the terminal nodes can choose the appropriate subset to minimize the positioning error and positioning consumption of the system resources,including the consumed energy,channel utilization and so on.Simulation results show that the proposed algorithm can achieve high precision positioning by selecting a small number of reference nodes,and improve the convergence speed of cooperative localization algorithm.Finally,considering that the algorithm selecting fixed number of cooperative nodes has some defects,a variable number of cooperative nodes selection algorithm based on EFI is proposed.The algorithm can select the least number of reference nodes to participate in the cooperation under the premise of satisfying the positioning accuracy specified by the system.Therefore,it can maximize the utilization of the channel resources and energy resources.Several other commonly used node selection algorithms are compared by simulation experiments,and the effectiveness of the proposed algorithm is proved.