基于增量学习神经模糊网络的机动目标跟踪

本文提出了基于增量学习神经模糊网络机动目标跟踪模型.当被跟踪目标发生机动时,该模型立刻检测到机动并对卡尔曼滤波器的自适应系统协方差进行精确估计,系统得到及时、正确的补偿.增量学习神经模糊网络能够随着环境变化,自动调整、找到最优的网络结构及参数,当发生机动时,总是能产生接近真实机动值的估计输出,从而提高跟踪性能及避免错误跟踪.仿真结果表明,该模型比传统的机动目标跟踪模型有更好的跟踪性能,并且该模型能动态的适应环境的变化,使系统更加实时,精确的跟踪机动目标.

Tracking Maneuvering Target Based on Neural Fuzzy Network with Incremental Leaning

The scheme of tracking maneuvering target based on neural fuzzy network with increased leaning is proposed.When tracked target maneuver occurs,the scheme can detected maneuver immediately to estimate the maneuver value accurately,then the tracking filter can be compensates correctly and duly by the estimated the maneuver value and system covariance.When environment changed,neural fuzzy network with increased leaning can find its optimal structure and parameters automatically to adopt to changed environment,and always produces estimated output very close to the true maneuver value that lead to good tracking performance and avoid miss-tracking when tracked target maneuver occurs.Results of simulation show that the performance is superior to the traditional schemes and the scheme can fit changed dynamic environment to track maneuver target accurately and duly.