genericsoup's profile - activity

Thanks! - this is exactly what I was trying to search for - was just looking in the wrong places!

Hi there,

I'm using a turtlebot to do some navigation stuff and have some high level logic that determines where the turtlebot should go (I have regions of interest on my map, and this higher level logic chooses a specific region and decides to move the robot to the centre of the region.).

One of the issues I ran into was when it picks the locations and passes the co-ordinates to the move_base server (which seems to be using global_planner) there is no guarantee that a valid plan can be found for those specific co-ordinates.

What I wanted to know is if there is any way to set a threshold for the accuracy of the planner - in a use case like this it is not actually important that the robot moves to exact co-ordinates specified - being roughly there is good enough.

I was thinking that I could probably just make it keep trying to generate plans by randomly picking new points within a certain radius until it succeeds, but this seemed a little inelegant.

Would appreciate any ideas/pointers on how to get around this issue :)

Edited question to include relevant code - unsure if you get a notification when I edit, hence the comment. Thanks :)

I'm trying to create a service that returns an array of a custom msg. However, when I call this using roservice call /myservice [input params...] I get the following error:

ERROR: service [/myservice] responded with an error: service cannot process request: handler returned wrong number of values: Invalid number of arguments, args should be ['parsed'] args are(operation: travel_to
target: a1
time_step: 0, operation:open ... )

Based on this I assumed that the type I was returning was wrong, so outputted the type just before the return statement in Python, however they seem to be what I expect them to be based on the documentation here

type: type 'list'
element type: class 'asp_model.msg._SubGoal.SubGoal'

.srv file

string goal
string entityA
string entityB
---
Subgoal[] parsed

string operation
string target
int16 time_step

Some additional info: I also tried simply returning a list of integers [1, 2, 3], which with the response type set to Int16[] which resulted in the same error. I'm running on ROS Hydro, Ubuntu 12.04, no other errors show up when running catkin_make or running the server

Any ideas on how to go about fixing this? I'm not sure if I'm completely misunderstanding the error message!

Edit1: Python on server, I've removed code indicated by the '...' as its some lengthy regex and string parsing stuff:

    def parse_answer(self, raw):
        parsed = []  # Use this to save list of SubGoals
        ....
        ....
        for step in anslist:
            ....
            ....
            if step:
                parsed.append(SubGoal(operation=operation[0], target=target[0], time_step=int(time_step)))

        return parsed

Cheers, I think I'm going to have to do this unless I find an alternate method over the next couple of days. Since it is pretty much a one off in this case, it shouldn't be too bad.

I've got some maps that I've generated using the gmapping/slam packages with a turtle bot. However I would prefer to be working within the simulated environment provided by Gazebo.

Is there a way I can get the pgm/yaml files generated by ROS packages such as gmapping into Gazebo? Ideally it would be loaded in as a heightmap, but even if I could just get the image in as a ground plane sort of thing that would be good.

Any ideas on how to go about doing this?

Thanks for the pointers guys, will have a go at this over the coming days.

I've being trying to get ROS + turtlebot (Hydro) installed on my mac (running osx 10.10). ROS installed happily. However I am running into a few issues with compiling the turtlebot specific packages from source ( http://wiki.ros.org/turtlebot/Tutoria... ).

When I run:

rosdep install --from-paths src -i -y

I get the following errors:

ERROR: the following packages/stacks could not have their rosdep keys resolved
to system dependencies:
kobuki_ftdi: No definition of [ftdi-eeprom] for OS [osx]
kobuki_qtestsuite: No definition of [pyqt4-dev-tools] for OS [osx]

As expected, continuing on with catkin-make doesn't work out too well.

How would I go about solving/getting arounds this? (I don't mind if its not a full blown answer, would at least like to be pointed in the right direction)

Also, just for reference, a somewhat similar question was asked quite a while back without that much success: http://answers.ros.org/question/89831...

I have used Git a bit before, but it just seems odd that I would have to push changes up just so I can get my files across to the robot? Unsure if you've used the Player framework before, but it allows you to deploy code from your workstation by passing in the IP address of the robot when launching

New to ROS here and I've been developing a simple package on my workstation. However, I want to run this package from my robot.

I could manually just copy the appropriate source files and rebuild it on the robot but that seems a little long winded - especially having to do this multiple times when developing.

I feel like I am missing something. What is the 'standard' way of doing this with ROS?