Abstract: Taking account of the problem that the simultaneous localization and mapping (Simultaneous Localization and Mapping, SLAM) method of underwater vehicle has low accuracy in the underwater environment with dense clutter and many map feature points, an improved random-finite-set to SLAM based on amplitude information has been proposed in this paper. The method uses the amplitude information of map features to estimate the map feature set and to obtain more accurate map features and clutter likelihood function, which improve the estimation accuracy of the feature map in SLAM process. This paper has researched the performance of the PHD-SLAM method with the addition of the amplitude information in the case of the known signal to noise ratio and the unknown signal to noise ratio. The results show that the proposed algorithm outperforms RB-PHD-SLAM in estimation of the number and location of map features and calculation efficiency.

Key words: Underwater vehicle, simultaneous localization and mapping, probability hypothesis density, autonomous navigation, amplitude information

摘要： 针对水下机器人同步定位与地图创建(Simultaneous Localization and Mapping，SLAM)方法在杂波密集及地图特征点多的水下环境中，存在地图特征点以及机器人自身位置估计精度较低的问题，提出一种基于幅值信息的随机有限集SLAM (AI-PHD-SLAM)方法。该方法利用地图特征量测的幅值信息获得更准确的地图特征和杂波的似然函数用以提高SLAM过程中对特征地图的估计精度。同时研究了在已知信噪比和未知信噪比的情况下，加入幅值信息后PHD-SLAM方法性能。通过仿真实验，将所提方法与RB-PHD-SLAM方法进行比较，结果表明该方法明显改善了地图特征位置和数目的估计精度并且有效降低了计算量。

关键词: 水下机器人, 同步定位与地图创建, 概率假设密度, 自主导航, 幅值信息