摘要
针对传统同步构图定位(SLAM)传感器具有数据量大、处理速度慢、实时性差的不足和基于扩展卡尔曼滤波的同步构图定位(EKF-SLAM)具有对水下无人航行器(UUV)位置估计精度低、甚至发散的缺陷,把带次优渐消因子的扩展卡尔曼滤波器应用到了导航系统中,提出了基于多元测距声呐(MRS)的UUV结构环境SFEKF-SLAM(suboptimal fadingextended Kalman filter-SLAM)方法.首先建立基于霍夫变换的水下MRS特征提取模型,设计了基于SFEKF-SLAM的UUV导航系统,利用该系统可以对UUV的状态进行预测,结合MRS信息可以对UUV周围结构环境进行状态更新.海试结果验证了基于霍夫变换的水下MRS模型能够有效提取环境特征,基于SFEKF-SLAM的UUV导航系统相对于常用的基于EKF-SLAM的UUV导航系统具有更高的定位精度,能够构建更加精确的港口堤岸地图.
Traditional simultaneous localization and mapping（SLAM）transducers have the shortcomings of a large amount of data,slow processing speed,and poor real time capability.Furthermore,the method based on extended Kalman filter（EKF）-SLAM has the deficiency of relatively low estimation precision for the location of unmanned underwater vehicles（UUV’s）.In consideration of these problems,a suboptimal fading extended Kalman filter was introduced into a UUV navigation system,and a method of suboptimal fading extended Kalman filter-SLAM（SFEKF-SLAM） for a UUV’s structural environment based on multi-ranging sonar（MRS）was presented.First,a feature extraction model of underwater multi-ranging sonar based on the Hough transform was established;then a UUV navigation system based on SFEKF-SLAM was designed.Applying the system,the UUV’s state can be estimated and its structural environment can be updated by the data of multi-ranging sonar.The sea-trial results show that the model of underwater multi-ranging sonar based on the Hough transform can extract environmental features efficiently,and a UUV navigation system based on SFEKF-SLAM has higher precision than one based on EKF-SLAM and can build the map of port embankment more precisely.It was also proven that the SFEKF-SLAM method for a UUV’s structural environment based on multi-ranging sonar has potential applications in underwater structural environments such as ports.
出处
《哈尔滨工程大学学报》
EI
CAS
CSCD
北大核心
2012年第8期1016-1021,共6页
Journal of Harbin Engineering University
基金
国家自然科学基金资助项目(50979017)
中央高校基本科研业务费专项基金资助项目(HEUCF21026)
2009年教育部高等学校博士学科点专项科研基金资助项目(20092304110008)
