基于MEP-UKF的组合导航滤波算法

针对目前应用于SINS／GPS组合导航系统中的扩展卡尔曼滤波(Extended Kalman
Filter, EKF)存在精度低、实时性差的缺点，提出一种基于模型误差预测(Model Error
Prediction, MEP)的Unscented 卡尔曼滤波(Unscented Kalman Filter,
UKF)。MEP-UKF滤波算法将惯性器件测量误差作为模型误差使用MEP进行实时预测的同时，采 用UKF估计载体的姿态、速度及位置等误差信息，并反馈给SINS系统来校正导航参数。MEP-U KF不仅克服了UKF必须假设惯性器件误差为高斯白噪声的局限性，而且降低了SINS／GPS组合 导航系统状态变量的维数，大大缩短了导航解算的时间。仿真结果表明，MEP-UKF的收敛速 度和滤波精度均明显优于EKF，更好地满足了工程应用中对导航精度和实时性的要求。

Integrated Navigation Filtering Algorithm Based on MEP-UKF

A method of combining model error prediction (MEP) and unscented Kalman filter (UKF) was put forward to solve the problems of low accuracy and poor real time, which consist in extended Kalman filter (EKF) applied in SINS/GPS integrated navigation system. MEP-UKF filtering algorithm considers the measurement error resulted from the inertial unit as the model error and predicts it in time through MEP. Then it uses UKF to estimate the vehicle error information, including attitude, velocity and position, which are fed back to SINS to correct the navigation parameters. Consequently, MEP-UKF not only avoids the limitation that UKF has to assume the measurement error as the Gaussian white noise, but also decreases the state dimension of SINS/GPS integrated navigation system, further shorten the navigation calculation time. Simulation results show that MEP-UKF is superior to EKF in convergence and precision, and satisfies the requirements of navigation precision and real time which is emphasized in project.