Motion Planning with base_local_planner
I'm a bit confused about how to use base_local_planner; I'm following this tutorial . Apologies if questions below seem silly; I'm a newcomer to ROS.
My robot configuration launch file looks like this:
<launch>
<node pkg="gazebo_plugins" type="hokuyo_node" name="hokuyo_node" output="screen">
<param name="sensor_param" value="param_value" />
</node>
< uncommented for display ; odom is a c++ program I wrote to listen to odometry information -->
<node pkg="learning_tf" type="odom" name="odom_node" output="screen">
<param name="odom_param" value="param_value" />
</node>
<node pkg="tf" type="tf_echo" name="tf_listener" output="screen" args="/base_link /map">
</node>
</launch>
My move_base file is the same as erratic_navigation's move_base.launch.
Questions:
- What are the sensor, odom and transform packages required for the robot configuration file? Are they listeners to the information published by move_base?
- The only odometry reading I get by looking at rostopic output is from erratic_odom. Is this enough?
- Is /base_link --> /map the right transformation for tf?
- What output should I expect after I run both launch files? Shouldn't I be getting a plan through which the robot moves, dumping output?
- Is hokuyo_node the right sensor?
Comments
Do you have a real robot that you want to port to ROS? Or do you want to test the navigation stack using a simulated robot in Gazebo (e.g., the Erratic or PR2)?
Martin Günther
(
2011-06-16 00:41:33 -0600
)edit
I'm using the Erratic simulation.
PKG
(
2011-06-16 03:18:54 -0600
)edit
Is there a particular reason for using the Erratic? Most Willow packages use the PR2 simulation and are better-supported than the Erratic packages. I'd start by playing around with the launch files recommended by Eitan below to get a feel for how everything plays together.
Martin Günther
(
2011-06-19 21:11:31 -0600
)editadd a comment
