Comments (9)
Thanks, how reproducible is this?
Does it fail every time? Or sometimes pass as mentioned in this issue #20
from fetch_gazebo.
I tried like 5 times, it was never successful. One thing is that I am using integrated GPU (Intel® HD Graphics 620).
Here is my environment setup:
Environment Setup
Make sure Ubuntu 18.04 and ROS Melodic are installed. We are using this version because it is also installed on the robot and it works.
(Currently only ROS Indigo is supported as Debian packages, but there are git branches for Ubuntu 18.04/ROS Melodic and both are installed on the onboard computer of the robot.)
Installing fetch_ros & fetch_gazebo from source
Install the dependencies for Fetch Robotics:
sudo apt-get update
sudo apt install ros-melodic-robot-controllers ros-melodic-costmap-2d ros-melodic-moveit ros-melodic-rgbd-launch
Create a catkin workspace: http://wiki.ros.org/catkin/Tutorials/create_a_workspace
In the carkin workspace src
directory, run:
git clone https://github.com/fetchrobotics/fetch_ros.git
git clone https://github.com/fetchrobotics/fetch_gazebo.git
In your carkin workspace root
directory, run
catkin_make
Test your installation in an empty world
roslaunch fetch_gazebo simulation.launch
Gazebo demo
Install dependencies:
sudo apt install ros-melodic-map-server ros-melodic-simple-grasping ros-melodic-amcl ros-melodic-move-base
demo:
roslaunch fetch_gazebo pickplace_playground.launch
roslaunch fetch_gazebo_demo pick_place_demo.launch
from fetch_gazebo.
I also made changes mentioned in PR #26.
from fetch_gazebo.
We have a new Intern: @ejichen who is working on updating our demos with the Fetch in Simulation and on the Real Robot.
She started this week and will work on a new cube picking demo.
I'll assign this to her to look into.
- I've noticed in Gazebo9, the lighting is much more accurate than it was in the past. We may need to update our simulation world to increase the lighting to improve the ability to detect the cube.
- There are a number of other tickets related to the control of the base/gripper/arm etc. I think we can consulate them all into one ticket to focus our efforts.
from fetch_gazebo.
Hi @umhan35,
I tried to reproduce the issue and the situation showed in the video might be similar to your problem.
On the first try, the Gazebo window shows that the robot flickered a little bit after it moved the arm (0:05 in the video). The movement shifted the robot, resulting in the mismatch of the robot frame and the planning scene frame (shows in the rviz window).
The second try didn't succeed because the previous movement made some changes to the position of the cube. However, if you let the simulation run for a while, you could see the pick and place demo succeed with the third try (since the script is a while loop of pick and place).
from fetch_gazebo.
@umhan35 I've made two URDF changes.
- Quick Fix: #37 (comment)
- Better Fix: #37 (comment)
@ejichen has been using 1. which is a quick fix, but I will make a PR with 2. which adds the caster wheels to the model
from fetch_gazebo.
Actually, I before committing the changes, I went to record before and after videos.
And surprisingly the fetch was able to grasp the cube, using the ccb5ad3 and the fetch-ros from debians https://github.com/fetchrobotics/fetch_ros/releases/tag/0.8.0.
I didn't use fetch_ros from source- the versions from source upgraded the fetch_ikfast_plugin ZebraDevs/fetch_ros#107 to fix deprecation warnings with the next version of moveit.
from fetch_gazebo.
I'd say the behavior is not consistent. I am not sure if it is a Gazebo issue.
from fetch_gazebo.
The behaviour isn’t expected to be consistent, it’s running MoveIt! with random sampling based motion planners.
But, if we make the URDF changes to stablize the base (the real Robot is very stable and doesn‘t wiggle or tilt when moving the arm) then it should behave closer to the real thing. In #37, I’ve posted videos of Indigo & Kinetic and the inertia/CoM values visually look different- despite not being changed.
from fetch_gazebo.
Related Issues (20)
- Controllers in Fetch Gazebo Melodic don't start correctly several times before the robot gets in a usable state HOT 1
- how to make fetch execute a pre-defined trajectory in gazebo. HOT 3
- how to use arm_with_torso_controller/follow_joint_trajectory controller?
- [BUG] Fetch Robot doesn't tries to pick up the cube HOT 2
- About the demo.py in fetch_gazebo_demo HOT 2
- [BUG] demo.launch will fail with Exception HOT 1
- [feature] Noetic support?
- [feature] how to multiple fetch control?
- [BUG] Terrible control on freight robot without casters
- [FetchIt!] Update the screw bin in simulation HOT 8
- [FetchIt!] SCHUNK machine model shoudn't have the cylinder inserted HOT 10
- [FetchIt!] [BUG] Gazebo world missing romanoff utility caddy HOT 5
- [FetchIt!] Final list of objects and simulation layout HOT 6
- [FetchIt!] [BUG] RobotStatePublisher for Schunk machine breaks robot's /odom transform HOT 8
- [FetchIt!] Modify chuck on Schunk machine HOT 5
- [Fetchit!] [BUG] Rotated gears "stand up" over time HOT 3
- fetch robot is dark in gazebo simulation playground HOT 4
- [BUG] Arm gets into unusable position when using moveit and moving the torso HOT 11
- [feature] Migrating to ROS2 HOT 16
- Adding namespace to Fetch Robot in Gazebo HOT 2
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from fetch_gazebo.