Robotics StackExchange | Archived questions

How to quickly and accurately control the position and orientation of the robot when navigation

How to quickly and accurately control the robot's position and the vehicle's body's orientation in a short time after reaching the target point when using a single line LiDAR and ROS1's Navigation Stack in conjunction with the gmapping algorithm for map building navigation. I hope to control the position within 2cm, and the pose error is within. My problem is that if the tolerance is set very small, the robot will spend an extremely long time searching for accurate position and pose. If the tolerance is too large, the position and pose control accuracy will not meet the requirements. I combine the "movebase" and "gmapping" algorithms through a launch file to achieve SLAM mapping while navigating <!-- 集成 SLAM 与 movebase --> <!-- 启动SLAM节点 --> <!-- 运行movebase节点 --> <include file="$(find nav)/launch/movebase.launch" />

Asked by Shuteng on 2023-05-28 09:18:20 UTC

Comments

Are you actively building a map, or using a pre-existing map? The distinction is important.

Asked by Mike Scheutzow on 2023-05-28 10:15:57 UTC

I am actively building a map, how I do this is when I send a goal to my car, it can plan a path and navigate to the goal, at the same time building the 2d environment.

Asked by Shuteng on 2023-05-29 00:12:27 UTC

Answers

There is no brief answer to this question. You need to do Map Building, Localization, Path Planning and Robot Control -- all with high quality. The outcome will be limited by how good your sensors are, and by how complex the map building task is.

If you are looking for a more specific answer, please edit this question and ask something more specific.

Asked by Mike Scheutzow on 2023-05-29 15:17:13 UTC

Comments