基于ROS的ICP-SLAM在嵌入式移动机器人上的实现与优化

Implementation and optimization of ROS based ICP-SLAM embedded robot

基于激光雷达的ICP-SLAM在机器人操作系统中无法直接运行,且通过PC实现算法时,没有高效运载计算机资源,产生资源浪费,同时降低了移动机器人系统的灵活性。通过在嵌入式系统中搭建机器人操作系统,在系统中移植MRPT函数库中的ICP-SLAM算法,对算法进行优化,调整算法运行的CPU占用率,使用基于激光雷达的移动机器人进行建图。结果表明机器人操作系统生成的栅格地图可以满足机器人自主导航的需求,SLAM效果更为直观,系统更加灵活,成本和硬件配置要求大大降低。证明优化后的算法更贴近实际的使用需求。

Laser-based ICP-SLAM cannot run directly in the robot operating system, and there is no efficient carrying of computer resources when implementing the algorithm on PC, resulting in waste of resources, while reducing the flexibility of mobile robot system. This paper built robot operating system on the embedded system, transplanted the ICP-SLAM algorithm of the MRPT library into the system, optimized and fixed the CPU occupancy rate of the algorithm, test the function of SLAM by the mobile robot based on the LIDAR. The results show that the grid map that robot operating system generated have met the requirements of robot autonomous navigation, and the effect of SLAM is more intuitive, besides, the system is more flexible, greatly reduce cost and hardware configuration requirements. It proves that the optimized algorithm is closer to the practical requirements.