This project contains several ROS packages, simple_mapping and ewok_*, and keyboard control.
- simple_mapping is used to perform simple point cloud stitching.
- ewok is a modified version of the ewok project, which contains ring buffer map, spline trajectory representation and fast replanning
- keyboard_control read keyboard value and publish control message.
You can clone the whole project to your catkin_ws/src and then catkin_make it.
cd <SOME_PATH>/catkin_ws/src
git clone https://github.com/ZbyLGsc/ring_buffer.git
cd ..
catkin_make
This will build both packages altogether.
This package performs simple mapping, with synchronized pointcloud and camera pose as input and transformed pointcloud as output. It publish pointcloud every 0.5 second.
Usage:
Firstly run:
rosrun simple_mapping simple_mapping
Then open rviz with the cloud.rviz file. If everything runs correctly, you should see pointcloud in rviz. Please make sure that topics of PointCloud and Pose are correctly set in simple_mapping.cpp
This package is a implement of RingBufferMap, which is always centralized around the camera(robot), and contains occupancy, signed distance and pointcloud normal information.
Usage:
roslaunch ewok_ring_buffer ring_buffer.launch
rviz -d ring_buffer.rviz
Run ring_buffer.launch, open ring_buffer.rviz and feed synchronized pointcloud and odometry as input. Please make sure that topics of PointCloud and Odometry are correctly set.
The ring_buffer map is published as visualization_msgs::Marker, as well as PointCloud2. The center of ring_buffer map is published as geometry_msgs::PoseStamped. They are synchronized using header.stamp, so you can use message_filters to subscribe them.
To transmit cloud2 using wifi connection, run:
roslaunch transmit.launch
rviz -d cloud.rviz
Currently the cloud is generated by server periodicaly. For cloud generated by ringbufffer, disabled the timer in socket_sever.cpp.
Note: this readme is not up-to-date. More information will be added.