ROS Catkin package to track people using ortree and cluster extraction from a fixed point.
Sensor Used for testing: Quanergy M8
Written and tested on: Ubuntu 16.04, ROS Kinetic
This package depends on the people_msgs package:
sudo apt install ros-kinetic-easy-markers ros-kinetic-kalman-filter
git clone https://github.com/kdhansen/people ~/catkin_ws/src
- Make your catkin workspace:
mkdir -p ~/catkin_ws/src
catkin_init_workspace ~/catkin_ws/src
catkin_make -C ~/catkin_ws
- Clone
kdhansen/people
to your workspace:git clone https://github.com/kdhansen/people ~/catkin_ws/src
- Clone this repo to your workspace:
git clone https://github.com/MyNameIsCosmo/lidar_body_tracking ~/catkin_ws/src
- Source your workspace
source ~/catkin_ws/devel/setup.sh
- Build your workspace
catkin_make -C ~/catkin_ws --pkg people_msgs
catkin_make -C ~/catkin_ws
- Initialize your LIDAR, or play your ROSBAG
rosbag play -l bagname.bag
- Launch
source ~/catkin_ws/devel/setup.sh
roslaunch lidar_body_tracking lidar_body_tracking.launch
- The URDF frame is QP308. You can change this in
/urdf/m8.launch.xacro
- You can change the default topics in the launch file
/launch/lidar_body_tracking.launch
- Dynamic Reconfigure for Link, Leaf size, min cluster, etc
- ROSParam for topics
- Clustering of indicies for person detection
- Output potential people to a topic
- Calculate person velocity
- Calculate certainty of person
- Estimate person height, size
- Track person based on previous location
- people_msgs/Person does not include orientation
- Control an RVIZ marker or something
- Support body tracking while lidar is moving, loop closure and Odom tracking.
- Comment and document code
- Object-oriented
- Clean-up code