基于惯性导航角度补偿的室内激光SLAM方法

Indoor laser SLAM method based on inertial navigation angle compensation

针对传统二维激光同步定位与地图构建(SLAM)方法无法适用于非平坦、复杂的非结构化环境的问题,提出了一种基于惯性导航角度补偿的激光SLAM方法。该方法通过基于高斯牛顿法的扫描匹配估计出机器人的位置,使得机器人位置的计算不再依赖里程计模型。同时,通过卡尔曼滤波器得到精确的角度信息并对激光数据进行补偿,提高构建地图的精度。实验平台为搭载二维激光雷达,基于ROS(robot operating system)系统的两轮差速移动机器人。实验结果表明,该算法创建的地图墙体角度误差由11°降到1°,匹配位置误差由±15 cm降到±5 cm。通过对地图精度和机器人位置的误差对比证实了该方法具有良好的场景适应性和实用性。

Aiming at the problem that the traditional two-dimensional laser simultaneous localization and mapping(SLAM)method cannot be applied to the non-flat and complex unstructured environment,this paper proposes a laser SLAM method based on inertial navigation angle compensation.The method estimates the position of the robot by scan matching based on the Gauss-Newton method,so that the calculation of the robot position is no longer dependent on the odometer model.At the same time,the Kalman filter is used to obtain accurate angle information and compensate the laser data to improve the accuracy to construct the map.The experimental platform is a two-wheel differential mobile robot based on robot operating system(ROS)system equipped with two-dimensional laser radar.The experimental results show that the angle error of the map wall created by the algorithm is reduced from 11°to 1°,and the matching position error is reduced from±15 cm to±5 cm.The comparison of map accuracy and robot position error proves that the method has good scene adaptability and practicability.