论文部分内容阅读
为了提高卡尔曼滤波估计精度,提出了一种基于回归神经网络补偿卡尔曼滤波器估计误差的方法。根据Elman网络与非线性ARMA模型工作原理的相似性,利用Elman网络做误差估计器,补偿卡尔曼滤波器的估计精度。实际舰船航行数据仿真测试表明,该方法有效可行。
In order to improve the accuracy of Kalman filter estimation, a method of compensating Kalman filter estimation error based on regression neural network is proposed. According to the similarity of the working principle of Elman network and nonlinear ARMA model, the Elman network is used as error estimator to compensate the estimation accuracy of Kalman filter. The actual ship sailing data simulation test shows that the method is effective and feasible.