基于异步估计自适应Kalman滤波系统设计

Design of Kalman filter system based on asynchronous estimation adaptive technology

一般的Kalman滤波器对系统噪声和测量噪声统计特性要求较为严格,当统计特性存在不确定性时,估计会造成较大的估计误差,甚至使滤波器发散。针对此问题,在基于新息的自适应估计和极大后验估计的基础上对Q和R进行异步估计;以某型号光纤惯性测量装置和GPS系统为背景的仿真实验结果表明,该方法能够有效地对Q和R存在不确定的组合导航系统的误差状态变量进行估计,并能较好地保证自适应滤波器的收敛性,进一步提高了估计精度。

For the integrated navigation system with ses statistics, applying the conventional Kalman filter to large state estimation error or even filter divergence. To uncertaintyness of both system and measurement noiestimate the error states of the system will lead to a estimate the states of the integrated navigation system with uncertain system and measurement noise covariances Q and R, an asynchronous estimation adaptive filter is presented, which is based on the features of inertial measurement unit（IMU） based on fiber optic gyroscopes （FOG） and GPS. In the adaptive filter, Q and R are estimated by the innovation-based adaptive estimation approach and the maximum posterior estimation approach respectively, and only Q or R is estimated at the same time. The simulation results show that by choosing the filter working moment, the adaptive filter can keep the filter convergence and estimates the error states of the integrated navigation system with uncertain Q and R elfectively.