摘要
An improved target tracking information differentiating system using the neural network to substitute for fuzzy rules is presented for the infrared-radar dual-mode guidance system. Since the neural network training based on the expert knowledge database is conducted off-line, the benefits for developing real-time tracking capabilities can be obtained. The network outputs the confidence degree denoted by the weight value of target information in the data fusion center according to two input variables of the measurement noise covariance and the tracking filter covariance. Simulation results show that the improved system can differentiate the target tracking information from the seeker fast and accurately.
双模导引头为制导系统提供的目标信息存在很大的不确定性,辨别目标跟踪信息的有效性是双模制导系统中的关键环节。为此本文提出了将描述跟踪目标信息不确定性的模糊推理规则转化为神经网络的方法,解决了常规模糊逻辑推理系统处理导引头跟踪信息耗时长的问题。试验结果表明,该方法能实现对导引头目标跟踪信息的快速、准确判别,故满足了雷达/红外双模制导系统的要求。
