基于方差补偿自适应卡尔曼滤波的南京铁路隧道沉降变形处理分析

Application of Self-adapting Kalman Filtering Basing on Variance Compensation in Data Processing of Deformation Monitoring of Nanjing Subway

当测量数据中存在粗差时,使用传统卡尔曼滤波对数据进行处理,状态向量的滤波估计值精度和可靠性会明显变差,甚至可能导致滤波发散而无法获得预测结果。通过使用方差补偿自适应卡尔曼滤波进行处理,结果表明能够减弱或消除粗差对数据的影响,从而提高模型的预测精度。结合工程实例分析表明,当观测数据中存在粗差时,使用方差补偿自适应卡尔曼滤波能有效地抵抗粗差的影响,提高数据处理的精度。

As the model error and dynamic noise of tradition Kalman filter are always ensured by experience and usually not exactly consistent with the real situation,filtering divergence might be caused leading to the failure of getting predication results.With self-adapting Kalman filtering basing on variance compensation,however,not only the prediction data can be filtered,but the model error and dynamic noise could be compensated,which hence improves the prediction accuracy of the model.Real examples show that self-adapting Kalman filtering basing on variance compensation has a better filtering and noise reduction effect and higher accuracy than traditional Kalman filter.