The paper proposes an algorithm based on the Multi-State Constraint Kalman Filter (MSCKF) algorithm to construct the map for robots special in the poor GPS signal environment. We can calculate the position of the robots with the data collected by inertial measurement unit and the features extracted by the camera with MSCKF algorithm in a tight couple way. The paper focuses on the way of optimizing the position because we adopt it to compute Kalman gain for updating the state of robots. In order to reduce the processing time, we design a novel fast Gauss–Newton MSCKF algorithm to complete the nonlinear optimization. Compared with the performance of conventional MSCKF algorithm, the novel fast-location algorithm can reduce the processing time with the kitti datasets.

中文翻译：

---

缩短机器人定位算法的处理时间

提出了一种基于多状态约束卡尔曼滤波（MSCKF）算法的算法，用于构造GPS信号环境较差的机器人专用地图。我们可以通过惯性测量单元收集的数据和相机通过MSCKF算法以紧密耦合的方式提取的特征来计算机器人的位置。本文重点介绍了优化位置的方法，因为我们将其用于计算卡尔曼增益以更新机器人的状态。为了减少处理时间，我们设计了一种新颖的快速高斯-牛顿MSCKF算法来完成非线性优化。与常规MSCKF算法的性能相比，新颖的快速定位算法可以减少kitti数据集的处理时间。