chargingbot

Husky + UR5e + Robotiq 85 Gripper in Simulation

IKFast Plugin, Yolov3 & SSD Real-time Object Detection, Lidar & Visual SLAM.

Tested in Ubuntu 18.04 ROS Melodic. If you use ROS Kinetic, modify $robotiq_gazebo/src/mimic_joint_plugin.cpp to suit Gazebo 7

Installation

Install the ROS controllers

sudo apt-get install ros*controller*

Install dependencies

sudo apt install ros-melodic-pointcloud-to-laserscan ros-melodic-rosbridge-server
sudo apt-get install ros-melodic-robot-localization
sudo apt-get install ros-melodic-interactive-marker-twist-server
sudo apt-get install ros-melodic-ar-track-alvar

Go to your workspace/src

git clone https://github.com/zimougao/chargingbot.git
sudo apt-get install ros-melodic-object-recognition-msgs
cd ..
catkin_make
source devel/setup.bash

Download the urdf file from Google Drive

Visit Google Drive: https://drive.google.com/drive/u/0/folders/1XtgpTrNwrcQiswtwLbZn0oNSMuTtQXNs
Go to File: Charging robot materials / Team files / Bingkun / parking_lot / car_with_port
Download meshes file.

$ roscd husky_ur5_world/

Replace meshes file with the new one.

Modify model path to your workspace

Change 2 lines of paths in test_05.world:

$ roscd husky_ur5_world/worlds/
$ gedit test_05.world

Replace "/home/zbk/catkin_ws/src/car_with_port/meshes/station01.dae"

"/home/zbk/catkin_ws/src/car_with_port/meshes/EVcharger.dae"

"/home/zbk/catkin_ws/src/car_with_port/meshes/parkinglot.dae"

"/home/zbk/catkin_ws/src/car_with_port/meshes/Mercedes_AMG.dae"

to "your path to husky_ur5_world pkg/meshes/station01.dae"

"your path to husky_ur5_world pkg/meshes/EVcharger.dae"

"your path to husky_ur5_world pkg/meshes/parkinglot.dae"

"your path to husky_ur5_world pkg/meshes/Mercedes_AMG.dae"

Starting Gazebo Simulation

Open a new terminal. The simulation is launched by:

$ roslaunch husky_ur5_world husky_artrack_world.launch

Starting MoveIt control

Open a new terminal. Launch the MoveIt control by:

$ roslaunch ur5ehuskyg_moveit demo.launch

You now should see the same posture in the Gazebo as the MoveIt environment. Set in Rviz the MotionPlanning/Planning Request/Planning Group to manipulator to drag the arm around. By Plan and Execute the arm will move to its new position.

Control Husky to trace ar track

Open a new terminal. Launch the QR code detector by:

$ roslaunch husky_ur5_world husky.launch

Open a new terminal. Launch the tracker by:

$ rosrun husky_ur5_world carrot.py

You should see the Husky tracing the QR code

Control your ur5 with robotiq gripper85

Before controlling the arm, please shut off the carrot.py after the robot reaches the point.
You can send commands to the joints or to the gripper.
The robot needs several steps to grab the gun.
Step one: Let the robot arm be prepared

$ roscd pick_test/scripts
$ chmod +x pick_up_prepere.py cartisan_ur5.py move_pre_pick.py
$ rosrun pick_test pick_up_prepere.py

!!!!!!!!!!!!!!!!attention!!!!!!! you need make sure trajectory is working complete. otherwise, you need controll the arm to position test1 using rivz planning section, then relaunch the pick_up_prepere.py

Step two: Shut off the husky.launch and use the camera on the robot arm to detect the tag

$ roslaunch husky_ur5_world manipulator.launch

Step three: Move the gripper to the proper position.

$ rosrun pick_test cartisan_ur5.py

in that steps, if controller failed, you need redo whole process again.

iWuKe / chargingbot