Research On Target Tracking And Resource Management Of Distributed Radar System

Distributed radar network system is a new type of radar system,which uses its widely distributed radar nodes and works together with each node in the way of network,thus forming an integrated system which can work together with multiple nodes and has high spatial reuse and system stability.Each radar node in the radar network system can not only act as an independent radar station to detect and track targets,but also cooperate with other sub-nodes through the scheduling management of the system center to cover and irradiate specific airspace or even the whole airspace,so as to realize more accurate and wider target search.The target tracking process of the system mainly includes the preliminary filtering of the measured data by each local radar node and the data fusion among several radar nodes.The selection of tracking algorithm inside the local node depends on the current system environment,while the selection of data fusion algorithm has an important impact on the final state estimation of the target.Both of the above are studied and discussed in this thesis.And the structure characteristics of distributed radar network makes it very dependent on radar system center for each local node scheduling and resource allocation,in the case of system resources have a certain limit,the radar system shall,according to the current task demand and the change of environment situation,choose the optimal adaptive node selection algorithm to optimize the node resources,and this is an important research content of this thesis.The research object of this thesis is the distributed radar network system.The theory and application of target tracking algorithm in radar network system are studied,and the appropriate data fusion structure is selected to carry out information fusion for multiple child nodes.Finally,some efficient and fast concepts and models of radar node resource optimization control algorithm are proposed,and a specific engineering environment is set for simulation verification.The details are as follows:1.Different tracking algorithm selection schemes are proposed for different system working environments.In pure working environment(single target with single measurement data),this thesis mainly studies how to convert nonlinear measurement data to linear data in spherical coordinate system,and based on this,three kinds of filtering algorithms for converting measurement data are proposed.For the purpose of engineering implementation,an algorithm combining the improved unbiased transformation with the uncorrelated transformation is proposed.The simulation software is used to experiment these filtering algorithms,and the performance of these algorithms is compared from the two indexes of estimation accuracy and complexity.In the complex environment(multiple measurements of a single target,in which there is clutter data),the theoretical concepts of the nearest neighbor algorithm and the data association algorithm are derived,and the simulation experiments of the two clutter tracking algorithms are carried out in the design scene,and the feasibility of the two algorithms is verified in the clutter working environment.2.Research is made on the multi-node information fusion algorithm which must be used in the radar network system.Firstly,the structure of centralized and distributed data fusion algorithms and their advantages and disadvantages are introduced.Then it mainly discusses the principle of distributed data fusion algorithm and public derivation,and establishes a simple radar network system model to carry out the simulation experiment of data fusion,and verifies the accuracy improvement effect of distributed data fusion algorithm on target state estimation.Finally,the distributed fusion algorithm is selected as the main architecture of multi-node information fusion in the distributed radar network system.3.By building the framework of the node resource management and control algorithm optimization model,the tracking accuracy based on the error covariance matrix is determined as the iterative target of the optimization algorithm.First,based on the exhaustive algorithm,an adaptive node selection algorithm based on predictive tracking accuracy is proposed,and two different algorithm models are established to meet the needs of different system tasks.Then,a minimum node selection algorithm based on the greedy algorithm is proposed.Compared with the exhaustive algorithm,the computational complexity is greatly reduced and a certain tracking accuracy is ensured.Finally,from the relationship between the signal-to-noise ratio of the target signal and the degree of attenuation in the radar equation,a node selection algorithm based on the degree of attenuation is derived.One is the heuristic algorithm that can reduce the initial node pool size of the later exhaustive method,and the other is The greedy algorithm and the degree of attenuation are used in combination to make the current node selection speed of the system reach the fastest and reach the lowest accuracy specified by the system.