摘要
动态定位的数据处理中广泛应用卡尔曼滤波,而卡尔曼滤波的应用要求动态模型(函数模型)和随机模型可靠和切合实际,但实际测量定位中难以保证观测对象的规则运动,因而容易出现模型误差。探讨在实际应用中存在模型误差时的卡尔曼滤波,研究动态定位时卡尔曼滤波的模型检测与校正,给出一种偏差分离估计方法。由于不存在状态增广,因而该方法计算效率高。最后以一数字仿真(模拟)实验论证方法的可行性。
Kalman filtering is widely used for data processing in Kinematic GPS Positioning, while the practical application of Kalman filtering requires the dynamic model （functional model） and the stochastic model to be reliable and accurate, yet it is difficult to maintain regular motion of the object in actual kinematic positioning, thus model biases are usually generated. In view of this problem of kinematic positioning, this thesis discusses Kalman filtering when model biases exist in practical applications, studies model bias detecting and correcting of Kalman filtering with kinematic positioning, and provides a departing estimation algorithm of model biases. Since expansion of state does not exist, this method is efficient and the values obtained are accurate. Lastly, a digital simulation is employed to prove the feasibility of this method.
出处
《测绘学报》
EI
CSCD
北大核心
2005年第4期294-299,共6页
Acta Geodaetica et Cartographica Sinica
基金
湖南省自然科学基金资助项目(02JJY2066)
关键词
卡尔曼滤波
预报残差
可靠性分析
Kalman filtering
predicted residual
reliability analysis
