水下机器人自适应卡尔曼滤波技术研究

水下机器人的位置和速度传感器受环境影响较大，数据滤波问题是运动控制的核心问题之一.给出了离散型卡尔曼滤波的基本方程，描述了卡尔曼滤波所具有的两个计算回路:增益计算回路和滤波计算回路.建立了水下机器人状态方程和量测方程，并在此基础上采用了自适应卡尔曼滤波方法对水下机器人的传感器数据进行了滤波分析.引入了渐消记忆指数加权方法.对时变噪声统计中，强调了新近数据的作用.避免了系统误差和量测误差统计特性的不准确对系统滤波效果的影响.滤波效果分析表明此方法能达到很好的滤波效果.

Research on underwater vehicle adaptive Kalman filter

AUV position and velocity sensors are affected by environment. Data filtering is one of important problems of AUV motion control. Discrete basic Kalman filter equation is given. Two loops of Kalman filter. plus loop and filter loop are described. AUV state equation and measuring equation are founded. Data from AUV sensors are disposed by adaptive Kalman filter with fading exponent. Fading memory exponent is introduced. New data are emphasized for time-varied data. This method avoids inaccuracy by system error and measuring error. Filter effect analysis proves that the method is effective.