Abstract Attitude determination for underwater vehicles in control and navigation applications has been always an important issue. MEMS-based accelerometer and gyroscope data can be usually used to determine the underwater vehicle attitude. However, the gyroscope sensors suffer from bias problem and the accelerometer data is not reliable in the presence of the external acceleration for the attitude determination. Therefore, data fusion for these two sensors is used to determine the attitude. But it should be noted that the common fusion algorithm does not perceive some special maneuvering accelerated motions and the use of accelerometer data causes a mistake in determining the Euler angles in these situations. Therefore, an algorithm which detects the external acceleration and prevents the use of accelerometer data in the fusion is extremely necessary for these situations. In this paper, an adaptive extended Kalman filter based on the quaternions is used by consideration of the effect of side accelerations to improve the detection of the maneuvering accelerated motions and reduce its weight in the fusion algorithm. An underwater motion simulator has been experimentally used and compared with the traditional methods to demonstrate the effectiveness of the proposed method.

中文翻译：

---

用于操纵水下航行器的加速运动的姿态确定改进

摘要 水下航行器在控制和导航应用中的姿态确定一直是一个重要的问题。基于MEMS的加速度计和陀螺仪数据通常可用于确定水下航行器的姿态。然而，陀螺仪传感器存在偏差问题，加速度计数据在存在用于姿态确定的外部加速度的情况下不可靠。因此，这两个传感器的数据融合用于确定姿态。但需要注意的是，普通融合算法无法感知一些特殊的机动加速运动，在这些情况下使用加速度计数据会导致确定欧拉角的错误。所以，对于这些情况，一种检测外部加速度并防止在融合中使用加速度计数据的算法是非常必要的。本文考虑到侧向加速度的影响，采用基于四元数的自适应扩展卡尔曼滤波器来改进对机动加速运动的检测，并在融合算法中降低其权重。实验使用了水下运动模拟器，并与传统方法进行了比较，以证明所提出方法的有效性。