Based On Chaos Sequence Data Transmission Of The Information Fusion Estimation

Information fusion system in the context of sensor networks during the datacollection, transmission and processing time will inevitably face the limitations of thenetwork constraints, such as random delays would lead to the result that when theoriginal orderly sampled data reaches the local processor or fusion center it willproduce the phenomenon of out-of-sequence and the out-of-sequence data transmissionwould lead to the result that the traditional Kalman filtering based on an orderly systemcannot be effectively and directly used in the out-of-sequence fusion system.Based on the above factors, this thesis carries out further research and analysis onthe topic of out-of-sequence data transmission information fusion in wireless sensornetworks under the framework of distributed, centralized and distributed twiceout-of-sequence. The main research contents and methods are as follows:First, in the centralized framework, based on the pseudo-measurement method,the thesis utilizes the algorithm of centralized observation weighted fusion for thenoise related multi-step-lag out-of-sequence transmission systems, and conducts adetailed analysis on the related noise. Moreover, in the case of nonlinear systems,multi-step-lag out-of-sequence transmission observation weighted fusion algorithm isderived based on the principle of extended Kalman filter, and analysis and simulationis conducted for the nonlinear system model of the two-station pure azimuth targettracking in the out-of-sequence transmission case.Second, in the distributed framework, based on the pseudo-measurement method,the thesis presents the method for single-sensor noise related multi-step-lagout-of-sequence transmission estimation, then the related noise is derived, and it usesCI algorithm and distributed weighted fusion algorithm for the case of multi-sensorfusion estimation. Moreover, in the nonlinear case, it analyzes the same out-of-sequence condition using the principle of extended Kalman filter, derives thefilter estimation of nonlinear single-sensor multi-step-lag out-of-sequence transmission,and uses distributed weighted fusion and CI fusion algorithm in the case ofmulti-sensor fusion estimation.Third, one framework for distributed twice out-of-sequence transmission ispresented. In this framework, the thesis introduces and analyzes two kinds of criteriafor the newest available local estimates (Newest Available Local Estimate, NALE), andusing the NALE criteria, preprocessing is conducted when the first timeout-of-sequence local estimation transmits to the fusion center to produce the secondtime out-of-sequence. Then it derives and calculates the distributed weighted fusionalgorithm and CI fusion algorithm in this framework for the results of preprocessing.A large number of simulation results show the correctness and validity of thetheoretical algorithm.