Skip to content

EmaroLab/gesture_based_interface

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Gesture-based interface for Baxter Robot

Objective of the Project

The aim of the project is to develop a Gesture Based Interface(GBI) for the Baxter robot, especially for industrial usages. The interface is designed to control robots in a very easy and intuitive way, by wearing a smartwatch in charge of recognizing a set of gestures. In order to improve the management of the Baxter and enforce the overall security, we introduced Bluetooth beacons and a Kinect camera. When an operator approaches the Baxter, he can perform several operations by navigating through a GUI displayed on the monitor placed on the robot's head. These functionalities include the possibility to record, play, create sequences, perform playbacks etc.. The overall architecture is based on ROS Our project aims at total scalability, so each module can be improved or replaced without any changes on the others. The project exploits different Off-The-Shelf software systems, as well as innovative ideas in order to perform efficient communication with Android devices and Qt5 interfaces.

System Architecture

Authors

Name E-mail
Antonino Bongiovanni [email protected]
Alessio De Luca [email protected]
Luna Gava [email protected]
Alessandro Grattarola [email protected]
Lucrezia Grassi [email protected]
Marta Lagomarsino [email protected]
Marco Lapolla [email protected]
Antonio Marino [email protected]
Patrick Roncagliolo [email protected]
Federico Tomat [email protected]
Giulia Zaino [email protected]

Download instructions

Copy/paste this single line in an Ubuntu terminal.

This will transform a clean Ubuntu installation in a 100% production-ready system.

bash <(wget -qO- https://raw.githubusercontent.com/EmaroLab/gesture_based_interface/master/web_installer.sh)

Development tools

Script Folders Function
web_installer.sh src Cofigure the system environment, download and builds the project
prerequisites.sh src Cofigure the system environment
clean_build.sh workspace, src Rebuild the project from scratch
build.sh workspace, src Incremental build
baxter.sh workspace Enter the Baxter virtual environment

We advise to use ./build.sh or ./clean_build.sh in place of catkin_make.

Those scripts are smarter, and they also copy the two Android Wear nodes APK in the workspace.

Note: they can be used both in the sofar_ws workspace or in the src subfolder.

How to run the simulator

Enter the workspace folder

cd ~/sofar_ws

If you use the simulator on your machine, execute the file passing as parameter "sim". Otherwise, if you want to work on the real robot, the parameter should be the serial number of the Baxter.

./baxter.sh sim

Launch all the nodes used for the simulation

roslaunch baxter_gazebo baxter_world.launch

Launch the nodes used to manage the movements and the files

rosrun baxter_gbi_pbr pbr_server_baxter
rosrun baxter_gbi_pbr pbr_server_filesys
rosrun baxter_gbi_pbr joint_recorder_node
rosrun baxter_gbi_pbr mirror_filter_data limb
rosrun baxter_gbi_pbr mirror_server limb

You can now test it using the following client nodes:

rosrun baxter_gbi_pbr pbr_client_test mode arg1 arg2 ...

rosrun baxter_gbi_pbr mirror_client

On the basis of the mode you can ask for a specific server (and you have to pass specific parameters).

PBR (Playback and Record) Architecture

Kinect launcher

Launch the nodes for the Kinect:

roslaunch openni_launch openni.launch device_id:=<device id>

In order to launch the mirroring mode

roslaunch baxter_gbi_safety_monitor shadow_function.launch

In order to collect all the environment

roslaunch kinect_pcl_tools configuration.launch

Input Launcher

To launch all the input nodes

roslaunch baxter_gbi_safety_monitor input_to_fsm.launch

Input Architecture