Cartographer SLAM with ouster Lidar and use_odometry = true

Hey forum members, greetings to you all! I am not very familiar with all aspects of ros & Its been around 1 month since I am struggling with 2 issues. Issue 1: It's related to cartographer with ouster lidar and Odom data from mavros. when I am launching cartographerros with useodometr = false, submaps are getting published without any error and then occupancygrid_node is also publishing /map but when I try to visualize things in RIVZ. The output is not stable it keeps rotating on its own axis even though the lidar is stable. I am not able to understand what is causing the moment of the points and ultimately the map itself is rotating.

Issue 2: when I am trying to enable Odom or I am trying to use the /imu data from mavros the cartographer starts raising a warning [ WARN] [1565342884.273398933]: W0809 09:28:04.000000 11280 orderedmultiqueue.cc:155] Queue waiting for data: (0, points2). This problem just arrises when FCU I try to integrate FCU along with Lidar data. Lidar data is getting published without any error. I am not able to understand why cartographer is not able to read those values if its some transformation error or anything else.

Please find launch, URDF & config file https://github.com/chetanchadha1/SLAM

Please help me how to integrate Odom with lidar in cartographer as without Odom data I am not able to visualize how SLAM will be possible. Thank you in advance

Asked by chetanchadha1234 on 2019-08-09 04:44:02 UTC

Comments

Answers

For the first point I need more information,

but I had problems with a bad calibrated IMU causing my Map to rotate. Can you check your sensor readings? Maybe one of your sensors is sending garbage values. But since you are not using the odometrie that actually should not cause the error. Can you record your data and use the validate rosbag?

Could you try to use only the ouster lidar?

In general you should also read more about SLAM. SLAM is possible even with lidar only , the key technology or better algorithm enabling SLAM is scan matching. You match the two point clouds after two scans and calculate the translation and delta orientation between them. Lidar only is often used for 2D-Slam algorithms.

So even without Odom and just Lidar and Imu you should be able to perform 3D slam.

Asked by RosNav on 2019-09-06 06:43:28 UTC

Comments

Thank you for your response. I will try to follow the instructions and get back if there will be an issue. For the first issue, it might beI MU issue I will try to change the imu and then get back.

Asked by chetanchadha1234 on 2019-09-17 06:08:58 UTC