Application of neural network in Infrared-Radar Dual-mode Guidance System

In this paper a novel target tracking information differentiating system with the capability of real time tracking and accurately identifying is developed in infrared-radar dual-mode guidance system. We take advantage of neural network´s well known capability of learning to perform the required classification without the assumption of probabilistic models for the input models to substitute the fuzzy rules in the information differentiating system. Since the neural network training based on expert knowledge database is conducted off-line, significant benefits of developing real time tracking capabilities are possible. Her supervising trained neural network is applied in the dual-mode guidance system, which outputs the confidence degree denoted as the weight value of target information in data fusion center according to the two input variables of measure noise and tracking error. The validity of this method is proved by simulation.