为了提高自动驾驶汽车环境感知与安全性,构建高精度的环境地图,提出了基于双目视觉与惯性测量单元(Inertial Measurement Unit, IMU)的多源异构信息融合同步定位与地图构建(Simultaneous Localization and Mapping, SLAM)算法。首先,针...为了提高自动驾驶汽车环境感知与安全性,构建高精度的环境地图,提出了基于双目视觉与惯性测量单元(Inertial Measurement Unit, IMU)的多源异构信息融合同步定位与地图构建(Simultaneous Localization and Mapping, SLAM)算法。首先,针对双目视觉与IMU信息融合的问题,采用紧耦合方法,结合双目视觉传感器的深度感知能力和IMU的快速运动捕捉能力,在系统初始化过程中,引入了一次最大后验估计对双目相机与IMU进行处理;然后,在后端优化中,采用基于滑动窗口的非线性优化算法求解最优位姿;最后,通过自动驾驶试验平台搭建了SLAM系统实物验证平台,设计完成了SLAM系统定位试验和相关性能验证试验。结果表明,双目视觉与IMU信息融合的SLAM系统相较于单目视觉惯性融合(VINS-Fusion)算法的定位精度可提升30.34%,在试验和实际场景中均表现出了有效性。设计的多源异构信息融合的SLAM系统能够显著提升定位精度,且在交通安全环境中具有良好的应用前景,对于提高自动驾驶系统的性能和安全性具有重要意义。展开更多
Aiming at that the successive test data set of the strapdown inertial measurement unit is always small,a Bayesian method is used to study its statistical characteristics.Its prior and posterior distributions are set u...Aiming at that the successive test data set of the strapdown inertial measurement unit is always small,a Bayesian method is used to study its statistical characteristics.Its prior and posterior distributions are set up by the method and the pretest,sample and population information.Some statistical inferences can be made based on the posterior distribution.It can reduce the statistical analysis error in the case of small sample set.展开更多
文摘为了提高自动驾驶汽车环境感知与安全性,构建高精度的环境地图,提出了基于双目视觉与惯性测量单元(Inertial Measurement Unit, IMU)的多源异构信息融合同步定位与地图构建(Simultaneous Localization and Mapping, SLAM)算法。首先,针对双目视觉与IMU信息融合的问题,采用紧耦合方法,结合双目视觉传感器的深度感知能力和IMU的快速运动捕捉能力,在系统初始化过程中,引入了一次最大后验估计对双目相机与IMU进行处理;然后,在后端优化中,采用基于滑动窗口的非线性优化算法求解最优位姿;最后,通过自动驾驶试验平台搭建了SLAM系统实物验证平台,设计完成了SLAM系统定位试验和相关性能验证试验。结果表明,双目视觉与IMU信息融合的SLAM系统相较于单目视觉惯性融合(VINS-Fusion)算法的定位精度可提升30.34%,在试验和实际场景中均表现出了有效性。设计的多源异构信息融合的SLAM系统能够显著提升定位精度,且在交通安全环境中具有良好的应用前景,对于提高自动驾驶系统的性能和安全性具有重要意义。
文摘Aiming at that the successive test data set of the strapdown inertial measurement unit is always small,a Bayesian method is used to study its statistical characteristics.Its prior and posterior distributions are set up by the method and the pretest,sample and population information.Some statistical inferences can be made based on the posterior distribution.It can reduce the statistical analysis error in the case of small sample set.
