intel_perceptual_server

Intel Perceptual Computing SDK server

Setup on Windows

Tested in Windows 7 64 bits.

Install and set up ROS

1. Instal

• Intel perceptual SDK and connect sensor
• Qt 4.8.5 (http://qt-project.org/downloads)

2. Install ROS for Windows (http://wiki.ros.org/hydro/Installation/Windows)
3. Setup catkin workspace (http://wiki.ros.org/win_ros/hydro/Msvc%20SDK)

• > mkdir C:/work
• > cd C:/work
• > winros_init_workspace overlay
• > cd C:/work/overlay
• > winros_init_build --underlays="C:/opt/ros/hydro/x86"
• > winros_make

4. Clone githun source into C:/work/overlay/src directory (with the installed git bash)

• git clone https://github.com/hiveground-ros-package/hiveground_messages.git
• git clone https://github.com/hiveground-ros-package/intel_perceptual_server.git

5. Build

• > cd C:/work/overlay
• > winros_make

6. Setup hostname and clear firewall

• c:\windows\system32\drivers\etc\hosts
• Add Linux machine hostname and IP and make sure that you can ping the Linux machin PC with name

Setup on Ubuntu

Tested in Ubuntu 12.04 64 bits (Kernel 3.8.0-30 with AMD HD5700).

Install and set up ROS

1. Install ROS Groovy (http://wiki.ros.org/groovy/Installation/Ubuntu)
2. Install the following packages

• $ sudo apt-get install ros-groovy-desktop-full
• $ sudo apt-get install ros-groovy-movit-full
• $ sudo apt-get install ros-groovy-moveit-pr2

4. Setup catkin workspace with overlay

• $ mkdir -p catkin_ros/src
• $ mkdir -p catkin_ws/src
• $ source /opt/ros/groovy/setup.bash
• $ cd catkin_ros/src/
• $ catkin_init_workspace
• $ cd ..
• $ catkin_make
• $ cd ~/catkin_ws/src/
• $ source ~/catkin_ros/devel/setup.bash
• $ catkin_init_workspace
• $ cd ..
• $ catkin_make

5. Setup environment

• $ echo "source ~/catkin_ws/devel/setup.bash" >> ~/.bashrc
• $ roscd <-- it should bring you to ~/catkin_ws/devel

6. Make sure about rosdep

• $ sudo rosdep init
• `$ rosdep update``

7. Setup hostname and clear firewall

• \etc\hosts
• Add Windows machine hostname and IP and make sure that you can ping the Windows machine with name

Build additional ROS dependencies

1. Clone github source to ~/catkin_ros/src/

• $ git clone https://github.com/hiveground-ros-package/realtime_tools.git
• $ git clone https://github.com/hiveground-ros-package/control_toolbox.git
• $ git clone https://github.com/hiveground-ros-package/ros_controllers.git
• $ git clone https://github.com/hiveground-ros-package/ros_control.git

2. Compile all packages

• $ cd ~/catkin_ros
• $ catkin_make

Build helping hand robot packages

1. Clone github source to ~/catkin_ws/src/

• $ git clone https://github.com/hiveground-ros-package/denso_bcap.git
• $ git clone https://github.com/hiveground-ros-package/denso_rc7m_controller.git
• $ git clone https://github.com/hiveground-ros-package/hiveground_controllers.git
• $ git clone https://github.com/hiveground-ros-package/hiveground_messages.git
• $ git clone https://github.com/hiveground-ros-package/hiveground_robots.git
• $ git clone https://github.com/hiveground-ros-package/hiveground_moveit.git
• $ git clone https://github.com/hiveground-ros-package/hiveground_rqt.git
• $ git clone https://github.com/hiveground-ros-package/MIDEM.git

2. Compile all packages

• $ cd ~/catkin_ws
• $ catkin_make

Working with gestures

Start a helping hand simulator on Linux

Open a new terminal for each line.

1. $ roscore
2. $ roslaunch midem_common transform_intel.launch
3. $ roslaunch midem_common bringup.launch xacro:=urdf/midem_with_solder_lead.xacro
4. $ roslaunch midem_with_solder_lead_moveit_config midem_with_solder_lead_moveit_planning_execution.launch
5. $ roslaunch midem_user_interaction user_interaction_intel.launch
6. $ rosrun rqt_interactive_probe rqt_interactive_probe

This should bring up rviz (ROS 3D visualization) screen and a setting windows "rqt_interactive_probe"

Start a intel_perception_server on Windows

Open a command window

1. > set ROS_MASTER_URI=http://linux_hostname:11311
2. > cd C:\work\overlay\devel\lib\intel_perceptual_server
3. > intel_perceptual_server.exe
4. Click start on Inter Percaptual Server windows.

This should start the camera with face tracking and hand gesture detection.

Controlling the robot

At this point, when moving hand in front of the camera, hand gesture should appear in the Inter Percaptual Server windows. The server will transmit hand positions to the Linux machine and they will appear as red dots in rviz screen. Pleaes see the video for usage details.