gp's profile - activity

Is your SICK node running and publishing laser scans on /scan topic? In your launch file you are nowhere initializing that node. Another thing is that the tutorial.launch file that you are using is for demo, assumes that tf between different frames like map->nav , nav->base_link etc are being published. If you run the demo rosbag file you will see that these tf are published in that rosbag file. If you want to run the hectorslam with SICK, tutorial.launch file will not work as it is. Have a look at this project hector_slam_example. Use hector_hokuyo.launch. replace hokuyo initialisation with SICK initialisation.

You are welcome. hey accept the answer as correct by clicking the tick mark :)

i have added launch files in my answer. These are working for me. Save these in any folder with same file names, edit "path_to_folder" in hector_hokuyo.launch and just launch the file 'hector_hokuyo.launch' it will initialise hokuyo, hector_slam node , rviz and also display the path in the map.

Hey mohan, scan rate for RPLIdar is very less to be suitable for hector slam. For modifications to RPLidar messages, have a look at this blog . Although it is about Gmapping and RPlidar, some modifications in RPLidar scan messages are suggested in the blog.

yes. Actually map can be generated using any of the SLAM algorithms or even manually generated in any image editor and fed to localization as png, jpeg or pgm files. Thisbook has a chapter "Chapter 9 Building Maps of the World" on this topic.

This is a robot localization problem. You can use AMCL package for this purpose. AMCL.

Sorry for very late answer. Have you set the parameter "pub_map_scanmatch_transform" and also set the parameter "tf_map_scanmatch_transform_frame_name" as scanmatcher_frame. These parameters must be set for hector_mapping node in order to get a transform between map and scanmatcher_frame. Have a look at the details of this parameters at link.

You dont need to modify entire package and rebuild it, just modifying the launch files and setting proper parameters will do. Have a look at this project hector_slam_example

you can accept the answer by clicking tick mark so that the question is closed.

you can simply search the package name on ros wiki or even google search will work you can search like "ros tf". you will get the package summary, installation guide tutorials etc on the wiki page . eg. for tf static_transform_publisher checkout this wiki page tf

The procedure with catkin_make is when you are installing some package from a source code. In that case, the source code and the compiled libraries etc will be in your catkin workspace. It is generally suggested to install package as binaries unless you want to make some changes in the code.

can you tell me the procedure you followed for installing gmapping? which your ros distribution? if you simply run following command, all the required packages for gmapping will be installed and be available in share folder of your ros directory.

sudo apt-get install ros-indigo-gmapping

if you installing using "apt-get" i.e. binaries of the package then they will be available in share folder of your ros distribution directory. eg. I have ros indigo, i have it in /opt/ros/indigo/share/gmapping.

if node "hector_trajectory_server" is running, you can use this command to save tiff map

rostopic pub syscommand std_msgs/String "savegeotiff"

for more details follow this tutorial last step is for saving map.

Did you get it?

If you just want to check hector slam you can follow this tutorial. After installing binaries you don't need to modify anything. You can straightaway use it.

then follow this tutorial setting your robot. For example you can check this project hector_slam_example. In hector_hokuyo.launch file, replace hokuyo part by kinect.

Instead of getting the source code and building it by yourself, you can download the binaries i.e. hector_slam package itself for your ROS distribution. You can use command

sudo apt-get install ros-your_ros_dist_name-hector-slam

for eg. in case of ros indigo you can use command

sudo apt-get install ros-indigo-hector-slam

this will download and install all the necessary packages for hector slam.

Use hector_hokuyo.launch file at launcher. Set arg "trajectory_source_frame_name" as "scanmatcher_frame" in the geotiff_mapper.launch file. It would be better to make a copy of "geotiff_mapper.launch" in your local folder use that.

Can you please share your launch files or elaborate what procedure you are following. In "geotiff_mapper.launch" file, you will need to set argument "trajectory_source_frame_name" to tf for which you want to visualize the path.

The hector_slam_launch tutorial.launch works with prerecorded rosbag files. In your case some modifications are required. have a look at this project hector_slam_example. In launch folder you will find hector_hokuyo.launch file, modify the hokuyo part with SICK sensor initialization.

In your launch file, you are adding two transforms first "base_footprint_to_laser_link" and also "base_frame_to_laser_link" doing this is not correct because in tf each leaf can have only one parent.

Have a look at this project hector_slam_example. You can use the launch file "hector_hokuyo.launch" in the launch folder. Modify this file to use SICK laser in place of hokuyo.

You have to publish a static transform between scan and base_frame to solve that error. Have a look at this project hector_slam_example . This project uses Hokuyo UTM-30LX with Hector_slam. You can use the launch file "hector_hokuyo.launch" in the launch folder. You will have to replace the hokuyo related parts with SICK i.e. initialising SICK sensor related node. Both of these sensors publish their laser scans on /scan topic so that should not be a problem.

Check out the parameter "map_update_interval" Gmapping wiki. This parameter corresponds to interval between map updates. Its default value is 5 sec.