中文摘要：

针对采用星载GPS（Global Position System）定位的LEO（Low Earth Orbiter）微纳卫星获得的位置速度数据不连续，而使用轨道动力学获得的连续的位置速度信息误差快速发散的问题，提出了一种基于非线性MPF（Model Predict Filter）的LEO微纳卫星定位方法，它采用非线性MPF预测的模型误差作为一步状态估计，同时使用GPS信息作为观测量，并与改进的扩展卡尔曼滤波组合，既可获得连续的卫星位置速度信息，又可获得相当于GPS单点定位的精度．仿真结果表明，此种方法可以有效地获得连续的微纳卫星位置速度信息，并且精度优于EKF（Extended Kalman Filter）．

英文摘要：

To solve two problems on micro-nano satellite that the dates of position and velocity is not consecutive for using global position system （GPS） individually and the errors of position and velocity radiate rapidly using orbital dynamics, a novel method of position determination for low earth orbiter（LEO） micro-nano satellite based on nonlinear model predict filter （MPF） was proposed. In this method, MPF predictd the errors of model which is because orbital dynamics modle is not precise as the first step of estimation, and the informations of GPS on satellite as observations, then combined with extended Kalman filter （EKF）which was improved to obtain the position and velocity not only consecutively but also precisely as GPS for single point position. Simulations show that, it is effectively to get position and velocity consecutively, and the pricition is better than EKF.