模态切换水下机器人微惯性导航分析  

Analysis on micro-inertial navigation of model-converted remotely operated vehicle

作　　者：张明[1] 刘慧婷[1] 眭翔[1] 宋振文[1] 刘海舰[1] 曾庆军[1] 

机构地区：[1]江苏科技大学电子信息学院,江苏镇江212003

出　　处：《计算机辅助工程》2015年第3期88-94,共7页Computer Aided Engineering

摘　　要：针对某模态切换远程遥控水下机器人(Model-Converted Remotely Operated Vehicle,MCROV),基于微机电系统(Micro-Electro-Mechanical Systems,MEMS)器件设计微惯性组合导航系统.该系统包括陀螺仪、加速度计、磁罗盘、深度传感器和微处理器等.采用互补滤波方法抑制陀螺漂移,基于四元数算法对陀螺仪积分,并以四元数为估计对象设计无损卡尔曼滤波算法,从而提高导航精度.分析梯度下降法原理,并研究其在四元数更新中的补偿作用.水池试验表明:互补滤波与无损卡尔曼滤波相结合的方法能够获得比较精确、稳定的MC-ROV导航信息.基于实测数据的算法仿真表明梯度下降法可以在一定程度上改善导航效果.Based on Micro-Electro-Mechanical Systems （MEMS） devices, an integrated micro-inertial navigation system is designed for a Model-Converted Remotely Operated Vehicle（MC-ROV）. The system includes gyroscope, aceelerometer, magnetic compass, depth sensor, micro-controller, and so on. The drifting of gyroscope is restrained by the complementary filter method, while the gyroscope is integrated by the quatemion algorithm and the quatemions are taken as the estimation objects to build an Unscented Kalman Filter （UKF） algorithm, which can improve the navigation accuracy. The gradient descent method is analyzed and its function in the updating of quaternion is studied. The pool test results show that the stable and accurate navigation information of MC-ROV can be obtained by the method that combines the complementary filter with Kalman filter. The algorithm simulations based on the measured data demonstrate that the gradient descent method can improve the navigation effects in a certain degree.

关 键 词：水下机器人 微惯性导航 互补滤波 无损卡尔曼滤波 四元数 梯度下降法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]