kaerman

说明：  实现卡尔曼滤波，可以看出，滤波过程是以不断地“预测—修正”的递推方式进行计算，先进行预测值计算，再根据观测值得到的新信息和kalman 增益（加权项），对预测值进行修正。由滤波值可以得到预测，又由预测可以得到滤波，其滤波和预测相互作用，并不要求存储任何观测数据，可以进行实时处理。(Kalman filtering, can be seen, the filtering process is constantly " forecast- Fixed" recursive manner calculated to predict the value of the first, and then according to the new information should be observed and the kalman gain (weighted items), on predictive value of the amendment. Value can be predicted by the filter, but also can be filtered by the forecast, and its interaction filtering and prediction, does not require the storage of any observational data, real-time processing.)