Research On Key Technologies Of Multi-Sensor Information Fusion Based On Air-Defence Radar Networks And Its Applications

In recent years, many international events, such as the Gulf War, the 911 terrorist attacks, the MH370 incident of Malaysia Airlines and so on, have exposed a mess of air-defence safety problems, including air-defence intelligence not shared, providing target property unclearly, tracking discontinuously, and situation assessment inaccurately, which seriously affect the war action and the fighting operation on terrorism. Bounded by the relationship of command, the weapon capability, the geographic shielding, the establishment, and the system, our army has also these problems to a certain extent, and therefore focused on the studies and constructions of air-defence radar network intelligence systems since 2010. The study in this paper is a trial for innovating and exploring, combined with the author’s work experiences, based on the state that he has engaged in radar intelligence system constructions and maintenance for decades. The subject in this paper is concerned with the key technologies of optimizing network deploying, the target recognition, the target tracking, and the threat assessment, of multi-sensor information fusion, based on air-defence radar network systems, and the issue of improving related technologies is discussed, and the whole studies are applied in practice. The studies develop multi-sensor information fusion technology employed in air-defence radar networks well, enhancing the integer performance of air-defence radar network systems.Aimming at realizing the air-defence radar network deploying optimization, improving the heterogeneous sensor information fusion target recognition, accomplishing passive sensors target tracking, and analyzing air-defence threat assessment, the paper carries out studies and practical applications on optimizing network deploying, the target recognition, the target tracking, and the threat assessment for air-defence radar network systems, by means of the theory and method of multi-sensor information fusion, including the ant colony optimization (ACO) algorithm, the fuzzy neural networks, the Gauss-Hermite filtering (GHF) algorithm, and the Bayesian networks (BN) algorithm.Firstly, the paper summarizes the basic definition and characteristic of multi-sensor information fusion, and describes the system models and main algorithms in information fusion. Heterogeneous sensor systems can realize the compensation of the information from different sensors, and avoid their respective disadvantages, but their fusing’s difficult, and lack of uniform mathematical methods. Passive sensors are not emitting signals themselves, and thus having the strong ability of survival and anti-stealth in the battle.Secondly, based on the air-defence radar network deploying scheme, and optimization method, we establish the air-defence radar network platform, in practice. The ACO algorithm based deploying method is studied, and according to this, the improved parameter selective preference evolution analysis method is proposed. Simulations are conducted for the both method, and the better deploying scheme is derived. Based on a fieldwork survey, we accomplish the practical establishing for the air-defence radar networks with simulation results, proving the validity and operability of the theoretical methods.Thirdly, the issue of air-defence radar network target recognition improving methods, based on the heterogeneous sensor information fusion, is studied. Constructing the heterogeneous sensor information fusion system, according to 3 heterogeneous sensors, is discussed. The improving of the clustering center select algorithm based on fuzzy neural networks in the feature fusion, and the Dempster-Shafer (D-S) evidence theory and conflict problem solution in the decision-level fusion are discussed. The reconnaissance for 4 batches of 4 airplanes is conducted, and the target characteristic parameters are derived, and the simulation experiment is carried out. And consequently, the improving of the target recognition for air-defence radar network systems is accomplished.Then, the bearings-only multi-target tracking fusion for air-defence radar networks, based on passive multi-sensors is discussed. The method of target tracking association updating for passive multi-radars, based on GHF algorithm, is analyzed under the bearings-only condition. The feasible partitions for the whole observations and targets are derived by conducting a computation, using the prior information of the targets. The reconnaissance for 2 batches of 2 airplanes is conducted for bearings-only passive tracking association, based on the air-defence radar network systems. Comparisons of the association results are drawn.And finally, the enemy plane threat class assessment in many cases is analyzed combined with air-defence radar network systems. The mathematical model using the technology of fuzzy dynamic Bayesian networks (FDBN) is constructed for the air-air combat threat assessment. The radar reconnaissance for 2 batches of 2 airplanes is conducted for the simulated air combat. The inference course based on the algorithm of threat assessment using fuzzy dynamic Bayesian networks (TAFDBN), and the sensitivity computation of FDBN algorithm are given by computer simulations. The studying results spread the applications of the fighting operation, of the air-defence radar network systems.