基于神经-模糊推理网络的机动目标跟踪方法

对目标机动的检测和准确跟踪，是目标跟踪研究中非常重要但难度较大的问题。在雷达和红外融合的跟踪系统，将神经-模糊推理网络与扩展Kalman滤波结合起来，形成闭环。即提取和目标机动有关的特征量送入神经-模糊推理网络，再估计目标机动输入，实现对机动目标的精确跟踪。仿真试验说明了本文所采用方法的有效性。

Tracking of Maneuvering Target Based on Neural Fuzzy Inference Network

The fast target maneuver detecting and highly accurate tracking are very important and rather difficult to be solved in target tracking study. For the radar/infrared image fused tracking system, an extend Kalman filter is combined with a neural fuzzy inference network to form a closed loop. That is, the features related to the target maneuver are extracted from radar and infrared sensors, and inputs are sent into the neural fuzzy inference network firstly, and then the targets maneuver inputs are estimated secondly, so that, the accurate tracking is achieved finally. The simulation results indicate that the new method is efficient for maneuvering target tracking.