中文摘要：

针对新型水下检测以及作业机器人在导航精度、体积方面的要求,设计了一套基于微机电器件的组合导航系统.系统为抑制陀螺漂移而采用了互补滤波方法,以四元数为估计对象设计卡尔曼滤波器.文中采用改进的自适应卡尔曼滤波器,增大新近数据的作用,减小陈旧数据的作用,避免滤波发散,提高导航精度.通过水池实验表明互补滤波和自适应卡尔曼滤波结合能够获得比较精确、稳定的水下机器人导航信息.同时,基于实测数据进行的算法仿真表明改进后的渐消记忆指数加权自适应卡尔曼滤波可以在一定程度上改善导航效果.

英文摘要：

Based on the requirements for the navigation precision and volume of detection and working of the novel underwater vehicle,a set of integrated inertial navigation system based on MEMS devices has been designed.Complementary filters have been used to restrain the drifts of gyroscope and the quaternions are selected as estimating values to design a Kalman Filter. This paper applies an improved adaptive Kalman Filter（ AKF）,stressing the effects of new data with adding weights to gradually leave old data behind to avoid divergence and to further improve the navigation effects. The pool experimental results indicates that the complementary filter and the Kalman filter can get stable and high-precision effects. Meanwhile,algorithm simulations based on the real measured data demonstrate that the improvement of fading memory AKF based on exponential weighting can improve the navigation results.