Applications: Leg Detector

How to run leg_detector

These instructions describe how to use the leg detector for the Turtlebot on Hydro. But, the process should be similar for other robots.

Install

Install the people stack: sudo apt-get install ros-hydro-people The leg_detector package will be installed to /opt/ros/hydro/share/leg_detector, you can also browse its source on Github.

The leg_detector package's leg_detector.launch file hard-codes a scan:=base_scan remapping argument. For the Turtlebot, the laser scan is just on the /scan topic, so you can get rid of this. Create new launch file called turtlebot.launch which gets rid of the scan:=base_scan argument.

Run

On the Turtlebot

Start the Turtlebot:

roslaunch turtlebot_bringup minimal.launch
roslaunch turtlebot_bringup 3dsensor.launch

Use the extended Kalman filter to get /odom_combined:

rosrun robot_pose_ekf robot_pose_ekf

Launch the leg detector:

roslaunch leg_detector turtlebot.launch

Visualize the data

On your workstation, run rviz. Set your global fixed frame to /odom_combined. Add a marker display, and set the update topic to /visualization_marker. Add a laser scan display, with the update topic set to /scan.

Troubleshooting

Verify laser scan

In rviz, verify that the laser scan shows up, and that it makes sense. Run rosnode info leg_detector and verify that it's subscribed to the right laser scan topic.

TF errors

As with most TF errors, you may want to synchronize the clocks on the robot and the workstation by running this command on both: sudo ntpupdate ntp.ubuntu.com.