基于极限学习机的修正当前统计模型跟踪算法

针对当前统计(Current Statistical,CS)模型由于先验知识的缺乏和其结构特点导致的跟踪精度降低的问题,通过使用极限学习机(Extreme Learning Machine,ELM)并根据目标历史状态信息,实时估计并修正CS模型的加速度估计值,提出了基于ELM的修正CS模型跟踪算法.ELM的训练结果表明...

Modified Current Statistical Model Tracking Algorithm Based on Extreme Learning Machine

Due to the lack of priori knowledge and the structural characteristics of the Current Statistical model, the trackingaccuracy of the Current Statistical model is reduced. A modified Current Statistical model tracking algorithm based on theExtreme Learning Machine is proposed, which uses Extreme Learning Machine to estimate and modify the acceleration esti.mation of the Current Statistical model in real time according to the historical state information of the target. The training re.sults of the Extreme Learning Machine show that the training speed is very fast and the generalization is very good. The Mon.te Carlo simulation results of test dataset and single trajectory show that the proposed algorithm can improve the accuracy ofposition and velocity estimation of the original algorithm by about 14% on the test dataset. In the single trajectory trackingexperiments, the ARMSE and PRMSE of position, velocity and acceleration are about one fourth of the Current Statisticalmodel, the maneuvering adaptability is good and the robustness is stronger, and the acceleration estimation is more stable.The Extreme Learning Machine has the advantages of simple structure, fast training speed and low added computationalcost, which has good practical application value.