改进的截断正态概率密度模型自适应滤波算法

应用当前统计模型跟踪机动目标时,模型参数机动频率和加速度极限值需要根据经验事先设定,在假设不准确的情况下,大大降低了模型的跟踪精度。针对此问题,基于截断正态概率密度模型,提出了一种新的参数自适应跟踪滤波算法。该模型算法通过使用距离函数来表征目标进行机动的强弱状况,采用指数型调整函数自适应调整目标的加速度极限值和机动频率,从而实现了对系统状态噪声和滤波增益的自适应调整,提高了机动模型与目标实际机动情况的匹配程度,提升了滤波器的跟踪性能。仿真结果表明:与常规ACS和TGPMKF算法相比,新算法在跟踪机动目标时,性能更优。

An Improved Adaptive Filtering Algorithm Based on Truncation Gauss Probability Model

Using current statistical model for maneuvering target tracking has a good effect. However, maneuvering frequency and ultimate acceleration are defaulted by human''s experience. When the given parameters are not accordant with actual situation, the capacity of tracking maneuvering target will decline. In view of the problem that the tracking performance of model is dependent on the prior parameters, this paper puts forward an adaptive tracking algorithm based on truncation gauss probability model for target tracking. In this model, the maneuvering situation of targets is characterized by the distance function, the status yawp and filtering gain of model is adaptively adjusted by using the exponential adjustment function to modulate maneuvering frequency and ultimate acceleration, which can improve the matching degree between maneuvering target model and the actual movement of the target. According to the simulation results, the capacity of tracking maneuvering target is improved, comparing TGPNMKF with ACS and TGPMKF.