2010/12/20 Antons Rebguns <arebgun@gmail.com>
Pablo,

Our group is working with erratic robots. As we talked the ua_erratic_player package is the current official erratic package. So we've been working with it for a while. However it has some lacks, e.g. we can't read sonar sensor values and we can't use neither the vision-ptz actuator nor the ranger-tilt-actuator.  
Unfortunately our erratics don't have a pan/tilt unit, sonar, or IR sensors that are connected directly to erratic's controller board, so couldn't really write/test any of those pieces. With that said, I will happily accept patches that expose this functionality to ROS from the player driver.

Perfect. The code is almost complete, we initially forked your ua_erratic_player package but definitely is better to patch your package.
 
<snip> 
- Regarding the vision-ptz actuator. What would be a good way to send commands? JointArray + robot_state_publisher? Maybe a look_at command topic of type Point3D??

The I would do it is the following. Since this is the lowest level driver, I would just expose each motor separately, i.e. if you have a pan and tilt motors, you should be able to control each one separately by just sending a float value that represents an angle of rotation. Things like look_at can be implemented on top of that.

Yep. It's the most straight forward method.
mmm but anyways sooner or latter we'll have to publish in a JointStateArray and connect it with a robot_state_publisher given that we'll need the vision_ptz frame to locate the vision device point cloud. This info can be published also by the erratic node. Do you agree? :-?
 
- For the ranger-tilt actuator I guess is better a topic of type (Header/float).

I am not sure that the header is all that useful here. I would assume the driver will just execute the command immediately, so there's no need for Header/stamp; frame wouldn't make much sense either. I would just go for plain float, same as pan/tilt unit.
 
Definitely.  I was thinking again in a "look at" method because initially I though that the ranger actuator was a ptz too.

One question more. Do you have' any erratic specification document where I may find precise measurements of the sonar devices arrangement? I need a more detailed urdf description to get all the sonar devices frames. This is needed to fill the Range message properly. The problem is that none of the specification documents of erratics that I find contain such information.

Regards.

--
Pablo Iņigo Blasco.
Computer Architecture Department. Universidad de Sevilla (Spain)
http://es.linkedin.com/in/pibgeus | http://geus.wordpress.com