eDO Control package

This package is a clone of eDO_control packages written by Yoan Mollard.

WARNING READ THIS

It has been modified to be compatible with eDO v3. If you have an edo version 2, please use the original package : https://github.com/ymollard/eDO_control

END WARNING

This package provides a simple ROS interface in order to execute trajectories with position PID control through the joint_trajectory_action_server and publishes the current /joint_states.

Advised ROS network configuration

Although other configurations are possible, this package intends to run on a deported workstation connected through Ethernet to a 6-axis e.Do on which an electric gripper is mounted. Other configurations (Wifi, no gripper, run on the internal Pi...) might need to tweak a bunch of parameters, as they have not been tested. We do not advise Wifi connection: since joint commands are sent over the network, it might result is jerky trajectories because of Wifi lags.

1. Local workstation network

Before anything, configure your workstation's IP address so that's in the 10.42.0.XXX network. Here we assume 10.42.0.1 so make sure you replace all 10.42.0.1 by you own IP if you choose another one.

2. EDo's raspi configuration over Ethernet

Allow your robot to communicate with your local workstation through Ethernet with this configuration step. Connect to the EDo's internal Raspberry Pi via SSH:

ssh [email protected]

Default password is raspberry. You should see an italian-speaking comandi tmux prompt with explanations about rostopic. If you don't, you have trouble in networking configuraiton, or the internal Raspberry Pi does not run.

We need to edit the ministarter script:

nano ~/ministarter

Locate these 2 ROS configuration lines:

export ROS_MASTER_URI=http://192.168.12.1:11311
export ROS_IP=192.168.12.1

and replace them by the IP address of the robot's Ethernet interface:

export ROS_MASTER_URI=http://10.42.0.49:11311
export ROS_IP=10.42.0.49

Press ctrl=X, y and then Enter to valide the filename and overwrite the file.

Type sudo reboot to reboot the robot and wait for it to be up again.

3. Local workstation ROS config

A setup script start.bash configures the environment variables for you every time you want to work with your EDo. It adds a yellow-coloured prefix to your prompt with the name of the right ROS master URI when enabled.

roscd edo_control
./start.bash

Check that everything is fine: If this command returns the message below, you're done for the ROS network configuration!

[http://10.42.0.49:11311] me@workstation :~$ rostopic echo /machine_state -n1
current_state: 0
opcode: 0

Now I want to start the ROS control interface

1. Calibration procedure is compulsory after any robot boot

 roslaunch edo_control calibrate.launch

You should first see a JOINT_UNCALIBRATED warning message and then the explanations for the calibration procedure. Follow the instructions every time you see Calibrating joint X, press left and right arrow keys to align the joint center marker and press Enter to switch to the next joint.

Go on with the entire calibration before continuing.

2. Bring up the action server and the joint state publisher

 roslaunch edo_control control.launch

This will bring up the following standard ROS interfaces:

• Topic /joint_states # Echoes joint states at about 90 Hz
• Action server /follow_joint_trajectory (e.g. to be used with eDO_moveit package)

Optionally, this interactive script offers more features for calibration, manual jogs, motions in cartesian space or returning to home position:

rosrun edo_control setup_edo.py

3. Open and close the gripper

Only the 2-finger electric gripper is supported in this package. You can publish true for opening the gripper (60mm range) and false for closing the gripper (0mm range) on the following topic:

rostopic pub /open_gripper std_msgs/Bool "data: true"