Research On Fusion Evaluation Method For Tactic Indexes Of Combined Guidance Weapon System

Taking combined guidance weapon system(CGWS) as the research object, this paper studies the fusion evaluation methods for three classic tactic indexes in small test samples, which are separately the success probability, the impact precision and the interdiction effectiveness. We take the tactic indexes as the object, analyze the influencing factors for the errors, establish the error models and then supply methods for the equivalent computation, error conversion and fusion evaluation.The main work and creativity of this paper are as follows:Researches on the fusion evaluation method for the success probability Based on the background of the active/passive radar aided navigation system (ANS), we study the target detection and recognition principles of the homing radar, and analyze the common characteristics and their influencing factors. We create the simulated data and imitate the prototype test by studying the simulation methods of the environment clutter and the target electromagnetic scattering.We establish the Logistic parameterization model for the success probability, fuse the heterogeneous test information and the prototype test data with the definition of the simulation credibility, estimate the coherent parameter by the regression analysis, and obtain the equivalent prediction method from the substitute test data to the prototype test result.We propose a fusion evaluation method of the target recognition probability in the minimum test samples. At first, we forecast the recognition probability in the prototype test by Logistic regression analysis of the influencing factors with substitute equivalent data. Then we simulate the prototype test by the target electromagnetic scattering simulation, the clutter simulation, and the interference simulation, obtain the testing value of the recognition probability. At last we take the weighted fusion evaluation for the prior information by the credibility research.Since the evaluating results of the success probability are not so good when linear model or nonlinear model is used to identify the weights of the prior distribution and subjective experiential information are usually not accurately chosen, we take the information entropy theory as a platform, convert the polygenetic heterogeneous test information into the prototype test data, identify the prior distribution of the success probability via the Bootstrap method, and then give the posterior distribution by the Bayesian method with prototype test information.Researches on the fusion evaluation method for the impact precisionBasing on the operating principle of the active/passive radar ANS, we analyze the error sources influencing the impact precision, establish the impact deviation model, and chiefly study the equivalent conversion methods for the inertial guidance instruments error and the homing radar guidance error.We propose an inertial guidance instruments error separation method by the sparse composition analysis, which can improve the coefficient correlation problem. In the method, we construct a weighted regularization model, discuss the criterion for selecting optimal model parameter, and design an adaptive iterative process.We propose an equivalent conversion method with process information for the homing radar guidance error, which validates the procedure error model with systematic error and random error characteristics by all observed information and provide a more robust evaluation result.We can get the substitute test conversion sample by the root variance compositing principle, and propose a Bayesian fusion evaluation method based on the prior information credibility. Firstly we take a consistency test between the substitute test conversion sample and the prototype test sample, define and compute the prior information credibility, then calculate the posterior distribution of the sample parameters, and finally get the evaluation result of the prototype test impact error. Researches on the fusion evaluation method for the interdiction effectivenessBased on the background of the submunition attacking the airport runway, we study the interdiction efficiency assessment model and the scheme optimization. Firstly we take the basal efficiency data, the bullets parameters, the target parameters, and the repair parameters as the initial variables, then simulate the impact distribution of the bullet cluster and the distribution of the attacking pits'size. In succession, we realize the computer simulations for the attacking procedure. Finally we get the assessment model of the destruction probability of the runway(DPR) and the interdiction time.Given the interdiction index level, we study the scheme design optimization problem. At first we choose the optimizing parameters, then establish the optimization model, and design two different algorithms for searching the optimization solution, which are the direct search algorithm and the genetic algorithm. In consider of the long calculation time when there are too many combination schemes in the direct search algorithm, we design a parallel algorithm in order to improve the calculation speed and efficiency.