论文部分内容阅读
近年用于水下滑翔器的低成本导航系统成为研究热点,导航器件的成本与精度之间的折中问题仍然是目前的难题。针对因使用低成本的导航元件而造成低精度位姿估计的问题,提出用于位姿估计的改进高斯混合粒子滤波(IGMPF)方法。用高斯混合模型来估计非高斯噪声,改进的粒子滤波进一步提高位姿估计精度。为了验证其效果,该方法应用于自主设计的水下滑翔器导航系统中并做了车载实验,实验结果表明所提IGMPF方法在实际应用中比传统的EKF和UKF表现更优,姿态角和位移误差比EKF和UKF减小了至少30%。
In recent years, underwater navigation glider for low-cost navigation system has become a research hotspot, navigation device cost and accuracy of the compromise between the problem is still the current problem. In order to solve the problem of low-precision attitude estimation caused by low-cost navigation components, an improved Gaussian mixture particle filter (IGMPF) method is proposed for pose estimation. The Gaussian mixture model is used to estimate the non-Gaussian noise. The improved particle filter further improves the pose estimation accuracy. In order to verify its effectiveness, the proposed method is applied to a self-designed underwater glider navigation system and a vehicle-mounted experiment is carried out. The experimental results show that the proposed IGMPF method performs better than the traditional EKF and UKF in practice, and the attitude angle and displacement The error is reduced by at least 30% over EKF and UKF.