SLAM(Simultaneous Localization and Mapping),即同时定位与地图构建,目前被广泛应用于机器人领域。SLAM算法使得机器人处于陌生环境时,能够通过自身搭载的传感器来感知环境信息并建立环境地图,并完成对自身位姿的计算,从而能够在未知...SLAM(Simultaneous Localization and Mapping),即同时定位与地图构建,目前被广泛应用于机器人领域。SLAM算法使得机器人处于陌生环境时,能够通过自身搭载的传感器来感知环境信息并建立环境地图,并完成对自身位姿的计算,从而能够在未知环境中进行移动。随着研究者们对SLAM问题的深入研究,SLAM领域相关成果已非常丰富,但是有关室内场景SLAM的论述还不够系统。通过对现有的关于SLAM算法发展成果的总结和对比,对室内SLAM进行了综合性的阐述。首先介绍了SLAM的技术现状和室内场景SLAM在不同传感器下的分类问题;其次介绍了SLAM的经典框架;然后根据相关传感器种类的不同,简要介绍了不同传感器下常见的SLAM算法的原理,同时讨论了传统室内SLAM算法中存在的诸多局限性问题,引出了基于多传感器融合技术的SLAM和基于深度学习技术的SLAM两个研究方向;最后介绍了SLAM的未来发展趋势和应用领域。展开更多
同步定位与建图(simultaneous localization and mapping, SLAM)技术作为智慧交通领域研究的热点,是无人驾驶车辆自主规划路径的关键。围绕SLAM技术相关传感器类型、定位、制图、多传感器融合四方面,从优缺点、适用范围、概率算法、地...同步定位与建图(simultaneous localization and mapping, SLAM)技术作为智慧交通领域研究的热点,是无人驾驶车辆自主规划路径的关键。围绕SLAM技术相关传感器类型、定位、制图、多传感器融合四方面,从优缺点、适用范围、概率算法、地图类型及融合方式出发,介绍SLAM技术实现过程中的各个环节,系统阐述了国内外相关的研究成果。基于多传感器融合SLAM,分析了目前常见的融合SLAM技术难题,对SLAM技术的未来发展趋势及实际工程应用做出展望。展开更多
In the present study, firstly, the unsteady cavitating flows around a hydrofoil are studied based on the flow visualization and detail velocity measurement, a high-speed video camera is used to visualize the flow stru...In the present study, firstly, the unsteady cavitating flows around a hydrofoil are studied based on the flow visualization and detail velocity measurement, a high-speed video camera is used to visualize the flow structures, and a particle image velocimetry （PIV） technique is applied to the measurement of the time-averaged and instantaneous velocity and vorticity fields. The results show that the unsteadiness of mass transfer process between the vapor and the two-phase regions is substantial, a self-oscillatory behavior of the whole sheet cavitation is obtained, with large length fluctuations and vapor cloud shedding, and also the cavitation structure depends on the interaction of the water-vapor mixture and the periodic vortex shedding. The main purpose of this experimental study is to offer information for validating computational models, and shed light on the unsteady multiphase transport process of cavitating flows. Furthermore, with an emphasis on the dynamics of the attached turbulent cavitating flows, a filter-based model （FBM） is derived from the k-6 two-equation model, a conditional averaging method aimed at improving unsteady simulation is applied to computation. In comparison to the standard k-ε model, overall, the filter-based model is shown to improve the predictive capability considerably.展开更多
文摘SLAM(Simultaneous Localization and Mapping),即同时定位与地图构建,目前被广泛应用于机器人领域。SLAM算法使得机器人处于陌生环境时,能够通过自身搭载的传感器来感知环境信息并建立环境地图,并完成对自身位姿的计算,从而能够在未知环境中进行移动。随着研究者们对SLAM问题的深入研究,SLAM领域相关成果已非常丰富,但是有关室内场景SLAM的论述还不够系统。通过对现有的关于SLAM算法发展成果的总结和对比,对室内SLAM进行了综合性的阐述。首先介绍了SLAM的技术现状和室内场景SLAM在不同传感器下的分类问题;其次介绍了SLAM的经典框架;然后根据相关传感器种类的不同,简要介绍了不同传感器下常见的SLAM算法的原理,同时讨论了传统室内SLAM算法中存在的诸多局限性问题,引出了基于多传感器融合技术的SLAM和基于深度学习技术的SLAM两个研究方向;最后介绍了SLAM的未来发展趋势和应用领域。
文摘同步定位与建图(simultaneous localization and mapping, SLAM)技术作为智慧交通领域研究的热点,是无人驾驶车辆自主规划路径的关键。围绕SLAM技术相关传感器类型、定位、制图、多传感器融合四方面,从优缺点、适用范围、概率算法、地图类型及融合方式出发,介绍SLAM技术实现过程中的各个环节,系统阐述了国内外相关的研究成果。基于多传感器融合SLAM,分析了目前常见的融合SLAM技术难题,对SLAM技术的未来发展趋势及实际工程应用做出展望。
基金supported by the National Natural Science Foundation of China (Grant Nos. 50679001 and 50979004)
文摘In the present study, firstly, the unsteady cavitating flows around a hydrofoil are studied based on the flow visualization and detail velocity measurement, a high-speed video camera is used to visualize the flow structures, and a particle image velocimetry （PIV） technique is applied to the measurement of the time-averaged and instantaneous velocity and vorticity fields. The results show that the unsteadiness of mass transfer process between the vapor and the two-phase regions is substantial, a self-oscillatory behavior of the whole sheet cavitation is obtained, with large length fluctuations and vapor cloud shedding, and also the cavitation structure depends on the interaction of the water-vapor mixture and the periodic vortex shedding. The main purpose of this experimental study is to offer information for validating computational models, and shed light on the unsteady multiphase transport process of cavitating flows. Furthermore, with an emphasis on the dynamics of the attached turbulent cavitating flows, a filter-based model （FBM） is derived from the k-6 two-equation model, a conditional averaging method aimed at improving unsteady simulation is applied to computation. In comparison to the standard k-ε model, overall, the filter-based model is shown to improve the predictive capability considerably.
