Fusion Assessment Method of Target Recognition Probability without Real Test Experiments

There is a strong military demand for assessing a certain tactical and technical performance index of a missile in battle without real test experiments,but the relative research is very little.Taking aim at assessing target recognition probability,a Logistic regression prediction model and a computer based simulation model were established to gain prior information;the fusion assessment method was performed by calculating each model's confidence,which is compatible with Bayes fusion assessment method having field subsample.Each model's confidence was computed by dividing the model into several parts and computing separately according to the physical background and the actual demand.The assessment sensitivity caused by the subjective parameters was also analyzed.